Search

Effective security surveillance is crucial in the railway sector to prevent security incidents, including vandalism, trespassing, and sabotage. This paper discusses the challenges of maintaining seamless surveillance over extensive railway infrastructure, considering both technological advances and the growing risks posed by terrorist attacks. Based on previous research, this paper discusses the limitations of current surveillance methods, particularly in managing information overload and false alarms that result from integrating multiple sensor technologies. To address these issues, we propose a new fusion model that utilises Probabilistic Occupancy Maps (POMs) and Bayesian fusion techniques. The fusion model is evaluated on a comprehensive dataset comprising three use cases with a total of eight real life critical scenarios. We show that, with this model, the detection accuracy can be increased while simultaneously reducing the false alarms in railway security surveillance systems. This way, our approach aims to enhance situational awareness and reduce false alarms, thereby improving the effectiveness of railway security measures. [ABSTRACT FROM AUTHOR]

In the domain of mobile robot navigation, conventional path-planning algorithms typically rely on predefined rules and prior map information, which exhibit significant limitations when confronting unknown, intricate environments. With the rapid evolution of artificial intelligence technology, deep reinforcement learning (DRL) algorithms have demonstrated considerable effectiveness across various application scenarios. In this investigation, we introduce a self-exploration and navigation approach based on a deep reinforcement learning framework, aimed at resolving the navigation challenges of mobile robots in unfamiliar environments. Firstly, we fuse data from the robot's onboard lidar sensors and camera and integrate odometer readings with target coordinates to establish the instantaneous state of the decision environment. Subsequently, a deep neural network processes these composite inputs to generate motion control strategies, which are then integrated into the local planning component of the robot's navigation stack. Finally, we employ an innovative heuristic function capable of synthesizing map information and global objectives to select the optimal local navigation points, thereby guiding the robot progressively toward its global target point. In practical experiments, our methodology demonstrates superior performance compared to similar navigation methods in complex, unknown environments devoid of predefined map information. [ABSTRACT FROM AUTHOR]

Smart and connected health (SCH) accelerates the development and integration of information science and engineering approaches to support the digital transformation of health and medicine in populated societies. Sensor data based human activity recognition (HAR) as an effective means of SCH is promising for healthcare monitoring and ambient assisted living. This paper focuses on multi-position sensor data fusion enabled HAR, and proposes a multi-channel deep learning framework. The main contributions include: (1) A multi-channel hybrid deep learning model (1DCNN-Att-BiLSTM) that merges a one-dimensional convolutional neural network, a bidirectional long short-term memory model, and an attention mechanism is proposed to significantly improve the recognition performance of HAR models by extracting and selecting local and global behavioral features in the spatial and temporal domains; (2) Publicly accessible datasets Shoaib AR, Shoaib SA, and HAPT are collected and processed to build multi-position sensor data pool for model's evaluation; (3) Thoroughly evaluating the classification performance of seven sensor data fusion patterns from different body positions within the parallel network structure of the multi-channel framework; (4) Comparing the recognition performance metrics of traditional machine learning models, deep learning models, and hybrid models with our proposed model. Extensive experiments demonstrate that our approach achieves competitive performance.

This study presents implementing and evaluating a computer vision platform to optimize warehouse inventory management. Integrating machine learning and computer vision technologies, this solution addresses critical challenges in inventory accuracy and operational efficiency, overcoming the limitations of traditional methods and pre-existing automated systems. The platform uses convolutional neural networks and open-source libraries such as TensorFlow and PyTorch to recognize and accurately classify products from images captured in real time. Practical implementation in a natural warehouse environment allowed the proposed platform to be compared with traditional systems, highlighting significant improvements, such as a 45% reduction in the time required for inventory counting and a 9% increase in inventory accuracy. Despite facing challenges such as staff resistance to change and technical limitations on image quality, these difficulties were overcome through effective change management strategies and algorithm improvements. The findings of this study identify the potential for computer vision technology to transform warehouse operations, offering a practical and adaptable solution for inventory management.

AbstractIntelligent Transportation Systems (ITS) are gaining momentum due to the advantages it possesses in congestion-free traffic, reduced probability of accidents, and economical transit thus reducing the transit time, saving human life, and helping the growth of the economy. Autonomous vehicles (AV) are an important part of ITS as these are the actual actuators of the ITS. In AVs, the perception system is particularly important as this provides vital information to the motion and planning system. Any error in perception of the environment by the AV will lead to the failure of the entire ITS. This article provides a thorough review of the recent technological advancements in perception systems. In this work, we have considered peer-reviewed journals and conference proceedings on AV from 2019 onwards. The primary focus is on motion prediction models, object detection, localization, sensor data fusion, sensors, autonomous driving (AD), communication technology and AI. This article also discusses the various challenges faced by perception systems in AD and how communication technology and Artificial Intelligence (AI) can help the perception system to overcome the existing challenges.

Horse riding, in its essence, is both an art and a captivating sport. In these events, riders showcase their refined skills and deep bond with their horses, most notably through changes in physical positioning. This research aims to accurately capture and analyze the spatial postures of riders in equestrian sports, exploring the dynamic differences between professional and amateur riders. To achieve this, we employed a multi-sensor data fusion approach based on human kinetics theories, enhanced by an extended Kalman filter. This method, combined with an optical tracking system, enabled us to intricately compare and analyze the 3D postures of riders during walking and trotting phases. Our research captured these nuanced postural shifts using an inertial sensor network, yielding nine-axis motion data of riders during their performance. The accuracy of our fusion technique was validated using the optical tracking system. By analyzing the motion data, we discerned posture variations among the riders of differing expertise levels, even when executing the same riding technique. The insights from this research offer a quantifiable metric for refining equestrian training and contribute to understanding the complex dynamics of rider–horse interaction. [ABSTRACT FROM AUTHOR]

In recent years, the continuous progress of wireless communication and sensor technology has enabled sensors to be better integrated into mobile devices. Therefore, sensor-based Human Activity Recognition (HAR) has attracted widespread attention among researchers, especially in the fields of wearable technology and ubiquitous computing. In these applications, mobile devices' built-in accelerometers and gyroscopes have been typically used for human activity recognition. However, devices such as smartphones were placed in users' pockets and not fixed to their bodies, and the resulting changes in the orientation of the sensors due to users' habits or external forces can lead to a decrease in the accuracy of activity recognition. Unfortunately, there is currently a lack of publicly available datasets specifically designed to address the issue of device angle change. The contributions of this study are as follows. First, we constructed a dataset with eight different sensor placement angles using accelerometers and gyroscopes as a prerequisite for the subsequent research. Second, we introduced the Madgwick algorithm to extract quaternion mode features and alleviate the impact of angle changes on recognition performance by fusing raw accelerometer data and quaternion mode features. The resulting study provides a comprehensive analysis. On the one hand, we fine-tuned ResNet and tested its stability on our dataset, achieving a recognition accuracy of 97.13%. We included two independent experiments, one for user-related scenarios and the other for user-independent scenarios. In addition, we validated our research results on two publicly available datasets, demonstrating that our method has good generalization performance. [ABSTRACT FROM AUTHOR]

The mobility sector is considered a major contributor to global greenhouse gas emissions and air pollution. As a result, many countries have initiated the transition from fossil fuel-powered to electrified powertrains. This transformation of the powertrain concept will lead to a rapid increase in the production of electric vehicles and, therefore, to a high demand for so-called traction batteries. As a production step of the traction batteries, a connection between the cell connector and the terminal of the battery cell has to be manufactured. For this purpose, laser beam welding is a reliable and efficient joining technique. In order to ensure continuous quality of the welding process during production and to detect defects in real time, reliable process monitoring is required. In this study, spectral and acoustic emissions during laser beam welding were recorded using a laser welding monitor and an optical microphone. For determining possible correlations between the signals and weld defects, various failure cases were generated by the systematic placement of disturbance elements. These elements included a contaminated surface, a gap between the cell connector and the battery cell, and a misalignment of the cell connector. Based on the recorded signals, statistical metrics were calculated. Finally, weld seams with and without defects were compared to assess the capability of both sensor systems for detecting the weld defects. [ABSTRACT FROM AUTHOR]

Petroleum products have contaminated groundwater with harmful organic compounds, such as benzene, toluene, ethylbenzene, and xylenes (BTEX). Collecting and analyzing polluted groundwater samples is expensive and undertaken infrequently. However, quick remedial action in case of unexpected events requires continuous monitoring. In‐situ water quality sensors (pH, EC, DO, ORP) may show correlations with the components of dissolved petroleum hydrocarbon (PHC) such as aromatics and non‐volatile mobile fractions. Correlations are prerequisite to ultimately develop real‐time prediction models. Since suitable field data sets are limited, we simulated the fate of hydrocarbons in groundwater under various realistic conditions using a reactive transport model as novel approach to explore when, where, and why correlations occur. A stationary oil source zone continuously dissolved at the top of a heterogeneous and shallow sandy aquifer over a two‐dimensional cross‐section. Our model considered transient conditions (fluctuating water table) and spatially uniform hydrogeochemical composition. We observed a strong correlation between PHCs and water quality sensors (rolling Spearman's correlation > |0.8|) at varying periods. These correlations are strongly affected by the location of observation wells, the aquifer's hydraulic conductivity, and the availability of calcite and oxide minerals, and other electron acceptors. DO and ORP are significant for the early detection of hydrocarbon contamination, whereas pH and EC are important features for the long‐term monitoring of hydrocarbons. Our findings lay the foundation for the subsequent development of a data analysis model to detect and estimate in real time PHC levels in groundwater using in‐situ water quality sensors. Plain Language Summary: Groundwater contamination by petroleum products, including benzene, toluene, ethylbenzene, and xylenes (BTEX), is a serious environmental concern. However, the collection and analysis of polluted groundwater samples are expensive and infrequently conducted. In the event of unexpected contamination, quick remedial action is crucial, which requires continuous monitoring of groundwater quality. We investigated whether in‐situ water quality sensors like pH, electrical conductivity (EC), dissolved oxygen (DO), and oxidation‐reduction potential (ORP) can predict the concentration of dissolved petroleum hydrocarbons (PHC) in groundwater. Understanding the correlations between these sensors and PHC levels is essential for developing real‐time prediction models. Since field data is scarce, we developed a reactive transport model that simulates the movement of PHC in groundwater under different conditions. Our model simulated a scenario where oil continuously dissolves at the top of a sandy aquifer. The correlation between PHC levels and water quality sensors varies depending on the location and timing of measurements. DO and ORP are particularly useful for detecting early signs of contamination, while pH and EC are important for long‐term monitoring. Our findings provide the foundation for developing data analysis models capable of real‐time detection and estimation of PHC levels in groundwater using water quality sensors. Key Points: In‐situ water quality sensors are highly correlated with petroleum hydrocarbons in groundwater when spatiotemporal variability is consideredDO and ORP are important for early detection, while pH and EC for long‐term prediction of hydrocarbon contamination in groundwaterThe aquifer's hydrogeology impacts hydrocarbon‐water quality correlations by influencing buffering capacity and flow velocity, among others [ABSTRACT FROM AUTHOR]

This paper proposes a novel data fusion technique for a wearable multi-sensory patch that integrates an accelerometer and a flexible resistive pressure sensor to accurately capture breathing patterns. It utilizes an accelerometer to detect breathing-related diaphragmatic motion and other body movements, and a flex sensor for muscle stretch detection. The proposed sensor data fusion technique combines inertial and pressure sensors to eliminate nonbreathing body motion-related artifacts, ensuring that the filtered signal exclusively conveys information pertaining to breathing. The fusion technique mitigates the limitations of relying solely on one sensor's data, providing a more robust and reliable solution for continuous breath monitoring in clinical and home environments. The sensing system was tested against gold-standard spirometry data from multiple participants for various breathing patterns. Experimental results demonstrate the effectiveness of the proposed approach in accurately monitoring breathing rates, even in the presence of nonbreathing-related body motion. The results also demonstrate that the multi-sensor patch presented in this paper can accurately distinguish between varying breathing patterns both at rest and during body movements. [ABSTRACT FROM AUTHOR]

Smart and connected health (SCH) accelerates the development and integration of information science and engineering approaches to support the digital transformation of health and medicine in populated societies. Sensor data based human activity recognition (HAR) as an effective means of SCH is promising for healthcare monitoring and ambient assisted living. This paper focuses on multi-position sensor data fusion enabled HAR, and proposes a multi-channel deep learning framework. The main contributions include: (1) A multi-channel hybrid deep learning model (1DCNN-Att-BiLSTM) that merges a one-dimensional convolutional neural network, a bidirectional long short-term memory model, and an attention mechanism is proposed to significantly improve the recognition performance of HAR models by extracting and selecting local and global behavioral features in the spatial and temporal domains; (2) Publicly accessible datasets Shoaib AR, Shoaib SA, and HAPT are collected and processed to build multi-position sensor data pool for model's evaluation; (3) Thoroughly evaluating the classification performance of seven sensor data fusion patterns from different body positions within the parallel network structure of the multi-channel framework; (4) Comparing the recognition performance metrics of traditional machine learning models, deep learning models, and hybrid models with our proposed model. Extensive experiments demonstrate that our approach achieves competitive performance. [ABSTRACT FROM AUTHOR]

Globally there is rising concern among postnatal women about the complications of Diastasis Recti Abdominis (DRA). DRA, the abdominal muscle divarication may susceptible to loss of functions of the abdominal wall. Rehabilitation exercises re-establish the connection between the abdominal and supporting muscles. The physiotherapists are currently unable to objectively detect the effectiveness of exercises in postnatal women to implement corrective measures. DRA subjects may exhibit diverse positions and movements during exercises emphasizing the need to consider movement patterns across various exercises. The muscle activation and movement patterns are assessed by the Electromyography (EMG) and Inertial measurement unit (IMU) Sensors. IMUs are essential sensors for evaluating exercise performance. The use of wearable IMU sensors enhances subject comfort during movements and facilitates the measurement of abdominal dynamics associated with various exercises. The preeminent objective of this work is to delineate the movement patterns of abdominal muscles during seven rehabilitation exercises specified by the physiotherapist for postnatal women. The trajectory of these exercises gives a clear view of the movement patterns. To predict the effective performance of exercises, modified complimentary filter (MCF) -based sensor fusion approach is utilized in finding the position information and trajectory of the movement patterns. The trajectory features provide insights of movement patterns based on the correctness class during each exercise. K-Nearest Neighbour classifier and Extreme Gradient (XG) Boost classifier are tested and yields an accuracy of 60% and 62%. The Binary Spiking classifier model, which leverages neuromorphic principles, is energy-efficient and capable at handling temporal dynamics. This Spiking model is used to evaluate the effectiveness of rehabilitation exercises, achieving an initial accuracy of 65%. To improve the model’s performance, IMU features are incorporated along with trajectory features, resulting in a higher accuracy of 85% compared to KNN and XG Boost classifier models.

LAPAN A2 is a Low Earth Orbit (LEO) satellite with the primary mission to monitor the area. In order to fulfil the mission’s objectives, the Satellite attitude must be controlled properly. One of the control processes under consideration is the output attitude estimation. Attitude estimation processes in satellites mostly use satellite sensor output transformation. The estimation process, however, cannot be carried out with only one sensor where the generated transformation does not represent the satellite actual attitude. Thus, a fusion of two or more attitude sensors is needed. However, measurement noise, such as the effect of the Earth's albedo on the sun sensor and the interference of the Earth's magnetic field on the magnetometer, is incorporated into the estimation process. Thus, a filtering process is needed. Hence, this study aims to investigate the utilization of the sensor output raw data fusion combined with the Kalman Filter algorithm to estimate the LAPAN-A2 satellite attitude in the form of Euler angle.

Indoor user positioning is a crucial problem in modern life. It has wide usage in health, security, smart homes, etc. Global positioning system (GPS) is used outdoors, and it does not work effectively in indoor areas since many things can degrade GPS positioning accuracy. All solutions for indoor areas aim to provide low-cost and high-accuracy positioning. In this study, a low-cost indoor positioning algorithm is developed. The fingerprint signal map of the building is measured with built-in digital sensors in smart devices. The measurements consist of Wi-Fi, Bluetooth Low Energy, and magnetic field signals called data fusion. During the positioning phase, the proposed model, called improved cosine similarity, uses the cosine similarity and information gain method. Digital magnetometers measure magnetic fields with different approaches. In the proposed method, Kalman filter is used to reduce noise magnetic field signals since this variety can give rise to mistaken positioning. To compare the effectiveness of the proposed method, it was compared to K-nearest neighbor, support vector machines, linear discriminant analysis, artificial neural networks, decision trees, N-nearest neighbor, and binned neighbor algorithm. Based on the experimental data, it was concluded that the proposed architecture achieved higher accuracy rates by reducing distortion. [ABSTRACT FROM AUTHOR]

Indoor fires pose significant threats in terms of casualties and economic losses globally. Thus, it is vital to accurately detect indoor fires at an early stage. To improve the accuracy of indoor fire detection for the resource-constrained embedded platform, an indoor fire detection method based on multi-sensor fusion and a lightweight convolutional neural network (CNN) is proposed. Firstly, the Savitzky–Golay (SG) filter is used to clean the three types of heterogeneous sensor data, then the cleaned sensor data are transformed by means of the Gramian Angular Field (GAF) method into matrices, which are finally integrated into a three-dimensional matrix. This preprocessing stage will preserve temporal dependency and enlarge the characteristics of time-series data. Therefore, we could reduce the number of blocks, channels and layers in the network, leading to a lightweight CNN for indoor fire detection. Furthermore, we use the Fire Dynamic Simulator (FDS) to simulate data for the training stage, enhancing the robustness of the network. The fire detection performance of the proposed method is verified through an experiment. It was found that the proposed method achieved an impressive accuracy of 99.1%, while the number of CNN parameters and the amount of computation is still small, which is more suitable for the resource-constrained embedded platform of an indoor fire detection system. [ABSTRACT FROM AUTHOR]

Effective security surveillance is crucial in the railway sector to prevent security incidents, including vandalism, trespassing, and sabotage. This paper discusses the challenges of maintaining seamless surveillance over extensive railway infrastructure, considering both technological advances and the growing risks posed by terrorist attacks. Based on previous research, this paper discusses the limitations of current surveillance methods, particularly in managing information overload and false alarms that result from integrating multiple sensor technologies. To address these issues, we propose a new fusion model that utilises Probabilistic Occupancy Maps (POMs) and Bayesian fusion techniques. The fusion model is evaluated on a comprehensive dataset comprising three use cases with a total of eight real life critical scenarios. We show that, with this model, the detection accuracy can be increased while simultaneously reducing the false alarms in railway security surveillance systems. This way, our approach aims to enhance situational awareness and reduce false alarms, thereby improving the effectiveness of railway security measures.

In the domain of mobile robot navigation, conventional path-planning algorithms typically rely on predefined rules and prior map information, which exhibit significant limitations when confronting unknown, intricate environments. With the rapid evolution of artificial intelligence technology, deep reinforcement learning (DRL) algorithms have demonstrated considerable effectiveness across various application scenarios. In this investigation, we introduce a self-exploration and navigation approach based on a deep reinforcement learning framework, aimed at resolving the navigation challenges of mobile robots in unfamiliar environments. Firstly, we fuse data from the robot’s onboard lidar sensors and camera and integrate odometer readings with target coordinates to establish the instantaneous state of the decision environment. Subsequently, a deep neural network processes these composite inputs to generate motion control strategies, which are then integrated into the local planning component of the robot’s navigation stack. Finally, we employ an innovative heuristic function capable of synthesizing map information and global objectives to select the optimal local navigation points, thereby guiding the robot progressively toward its global target point. In practical experiments, our methodology demonstrates superior performance compared to similar navigation methods in complex, unknown environments devoid of predefined map information.

We developed and applied a novel approach for shape agnostic detection of multiscale flaws in laser powder bed fusion (LPBF) additive manufacturing using heterogenous in-situ sensor data. Flaws in LPBF range from porosity at the micro-scale (< 100 µm), layer related inconsistencies at the meso-scale (100 µm to 1 mm) and geometry-related flaws at the macroscale (> 1 mm). Existing data-driven models are primarily focused on detecting a specific type of LPBF flaw using signals from one type of sensor. Such approaches, which are trained on data from simple cuboid and cylindrical-shaped coupons, have met limited success when used for detecting multiscale flaws in complex LPBF parts. The objective of this work is to develop a heterogenous sensor data fusion approach capable of detecting multiscale flaws across different LPBF part geometries and build conditions. Accordingly, data from an infrared camera, spatter imaging camera, and optical powder bed imaging camera were acquired across separate builds with differing part geometries and orientations (Inconel 718). Spectral graph-based process signatures were extracted from this heterogeneous thermo-optical sensor data and used as inputs to simple machine learning models. The approach detected porosity, layer-level distortion, and geometry-related flaws with statistical fidelity exceeding 93% (F-score).

This paper presents a leader–follower mobile robot control approach using onboard sensors. The follower robot is equipped with an Intel RealSense camera mounted on a rotating platform. Camera observations and ArUco markers are used to localize the robots to each other and relative to the workspace. The rotating platform allows the expansion of the perception range. As a result, the robot can use observations that are not within the camera's field of view at the same time in the localization process. The decision-making process associated with the control of camera rotation is implemented using behavior trees. In addition, measurements from encoders and IMUs are used to improve the quality of localization. Data fusion is performed using the EKF filter and allows the user to determine the robot's poses. A 3D-printed cuboidal tower is added to the leader robot with four ArUco markers located on its sides. Fiducial landmarks are placed on vertical surfaces in the workspace to improve the localization process. The experiments were performed to verify the effectiveness of the presented control algorithm. The robot operating system (ROS) was installed on both robots. [ABSTRACT FROM AUTHOR]

LAPAN A2 is a Low Earth Orbit (LEO) satellite with the primary mission to monitor the area. In order to fulfil the mission’s objectives, the Satellite attitude must be controlled properly. One of the control processes under consideration is the output attitude estimation. Attitude estimation processes in satellites mostly use satellite sensor output transformation. The estimation process, however, cannot be carried out with only one sensor where the generated transformation does not represent the satellite actual attitude. Thus, a fusion of two or more attitude sensors is needed. However, measurement noise, such as the effect of the Earth's albedo on the sun sensor and the interference of the Earth's magnetic field on the magnetometer, is incorporated into the estimation process. Thus, a filtering process is needed. Hence, this study aims to investigate the utilization of the sensor output raw data fusion combined with the Kalman Filter algorithm to estimate the LAPAN-A2 satellite attitude in the form of Euler angle. [ABSTRACT FROM AUTHOR]

The application of Global Navigation Satellite Systems (GNSSs) in the intelligent railway systems is rapidly developing all over the world. With the GNSS-based train positioning and moving state perception, the autonomy and flexibility of a novel train control system can be greatly enhanced over the existing solutions relying on the track-side facilities. Considering the safety critical features of the railway signaling applications, the GNSS stand-alone mode may not be sufficient to satisfy the practical requirements. In this paper, the key technologies for applying GNSS in novel train-centric railway signaling systems are investigated, including the multi-sensor data fusion, Virtual Balise (VB) capturing and messaging, train integrity monitoring and system performance evaluation. According to the practical characteristics of the novel train control system under the moving block mode, the details of the key technologies are introduced. Field demonstration results of a novel train control system using the presented technologies under the practical railway operation conditions are presented to illustrate the achievable performance feature of autonomous train state perception using BeiDou Navigation Satellite System (BDS) and related solutions. It reveals the great potentials of these key technologies in the next generation train control system and other GNSS-based railway implementations.

In recent years, the continuous progress of wireless communication and sensor technology has enabled sensors to be better integrated into mobile devices. Therefore, sensor-based Human Activity Recognition (HAR) has attracted widespread attention among researchers, especially in the fields of wearable technology and ubiquitous computing. In these applications, mobile devices’ built-in accelerometers and gyroscopes have been typically used for human activity recognition. However, devices such as smartphones were placed in users’ pockets and not fixed to their bodies, and the resulting changes in the orientation of the sensors due to users’ habits or external forces can lead to a decrease in the accuracy of activity recognition. Unfortunately, there is currently a lack of publicly available datasets specifically designed to address the issue of device angle change. The contributions of this study are as follows. First, we constructed a dataset with eight different sensor placement angles using accelerometers and gyroscopes as a prerequisite for the subsequent research. Second, we introduced the Madgwick algorithm to extract quaternion mode features and alleviate the impact of angle changes on recognition performance by fusing raw accelerometer data and quaternion mode features. The resulting study provides a comprehensive analysis. On the one hand, we fine-tuned ResNet and tested its stability on our dataset, achieving a recognition accuracy of 97.13%. We included two independent experiments, one for user-related scenarios and the other for user-independent scenarios. In addition, we validated our research results on two publicly available datasets, demonstrating that our method has good generalization performance.

This paper proposes a novel data fusion technique for a wearable multi-sensory patch that integrates an accelerometer and a flexible resistive pressure sensor to accurately capture breathing patterns. It utilizes an accelerometer to detect breathing-related diaphragmatic motion and other body movements, and a flex sensor for muscle stretch detection. The proposed sensor data fusion technique combines inertial and pressure sensors to eliminate nonbreathing body motion-related artifacts, ensuring that the filtered signal exclusively conveys information pertaining to breathing. The fusion technique mitigates the limitations of relying solely on one sensor’s data, providing a more robust and reliable solution for continuous breath monitoring in clinical and home environments. The sensing system was tested against gold-standard spirometry data from multiple participants for various breathing patterns. Experimental results demonstrate the effectiveness of the proposed approach in accurately monitoring breathing rates, even in the presence of nonbreathing-related body motion. The results also demonstrate that the multi-sensor patch presented in this paper can accurately distinguish between varying breathing patterns both at rest and during body movements.

In the age of big data availability, data-driven techniques have been proposed recently to compute the time evolution of spatio-temporal dynamics. Depending on the required a priori knowledge about the underlying processes, a spectrum of black-box end-to-end learning approaches, physics-informed neural networks, and data-informed discrepancy modeling approaches can be identified. In this work, we propose a purely data-driven approach that uses fully convolutional neural networks to learn spatio-temporal dynamics directly from parameterized datasets of linear spatio-temporal processes. The parameterization allows for data fusion of field quantities, domain shapes, and boundary conditions in the proposed U p -Net architecture. Multi-domain U p -Net models, therefore, can generalize to different scenes, initial conditions, domain shapes, and domain sizes without requiring re-training or physical priors. Numerical experiments conducted on a universal and two-dimensional wave equation and the transient heat equation for validation purposes show that the proposed U p -Net outperforms classical U-Net and conventional encoder–decoder architectures of the same complexity. Owing to the scene parameterization, the U p -Net models learn to predict refraction and reflections arising from domain inhomogeneities and boundaries. Generalization properties of the model outside the physical training parameter distributions and for unseen domain shapes are analyzed. The deep learning flow map models are employed for long-term predictions in a recursive time-stepping scheme, indicating the potential for data-driven forecasting tasks. This work is accompanied by an open-sourced code. [ABSTRACT FROM AUTHOR]

This paper concerns the novel concept of an Interactive Dynamic Intelligent Virtual Sensor (IDIVS), which extends virtual/soft sensors towards making use of user input through interactive learning (IML) and transfer learning. In research, many studies can be found on using machine learning in this domain, but not much on using IML. This paper contributes by highlighting how this can be done and the associated positive potential effects and challenges. An IDIVS provides a sensor-like output and achieves the output through the data fusion of sensor values or from the output values of other IDIVSs. We focus on settings where people are present in different roles: from basic service users in the environment being sensed to interactive service users supporting the learning of the IDIVS, as well as configurators of the IDIVS and explicit IDIVS teachers. The IDIVS aims at managing situations where sensors may disappear and reappear and be of heterogeneous types. We refer to and recap the major findings from related experiments and validation in complementing work. Further, we point at several application areas: smart building, smart mobility, smart learning, and smart health. The information properties and capabilities needed in the IDIVS, with extensions towards information security, are introduced and discussed. [ABSTRACT FROM AUTHOR]

Air quality monitoring is a process that determines the number of pollutants in the air, one of which is indoor air quality. The Fuzzy Indoor Air Quality Index was developed in this research. It is a method for determining the indoor air quality index using sensor fusion and fuzzy logic. By combining several different time series determinants of air quality, a fuzzy logic-based sensor fusion method is used to build a knowledge base about indoor air quality levels. Without the use of complicated calculation models, fuzzy logic-based fusion will make it easier to determine indoor air quality levels based on various sensor parameters. The input for fuzzy-based data fusion is obtained from the ARIMA method with Kalman Filter's air quality parameter values estimation. The application of ARIMA with a Kalman Filter was used to improve the accuracy of indoor air quality estimation in this study. ARIMA(3,1,3) had a MAPE of 0.1 percent on the CO2 dataset, and ARIMA(1,0,1) had a MAPE of 0.63 percent on the TVOC dataset based on approximately three experimental days. ARIMA (3,1,3) estimation with a Kalman Filter results in a MAPE of 0.03 percent for the CO2 dataset and a MAPE of 0.24 percent for ARIMA(1,0,1) Kalman Filter estimation on TVOC dataset. As a result, the Fuzzy Indoor Air Quality Index (FIAQI) developed in this research reasonably estimates indoor air quality. This can be seen by examining the percentage of estimation errors obtained from the experiment.

We developed and applied a novel approach for shape agnostic detection of multiscale flaws in laser powder bed fusion (LPBF) additive manufacturing using heterogenous in-situ sensor data. Flaws in LPBF range from porosity at the micro-scale (< 100 µm), layer related inconsistencies at the meso-scale (100 µm to 1 mm) and geometry-related flaws at the macroscale (> 1 mm). Existing data-driven models are primarily focused on detecting a specific type of LPBF flaw using signals from one type of sensor. Such approaches, which are trained on data from simple cuboid and cylindrical-shaped coupons, have met limited success when used for detecting multiscale flaws in complex LPBF parts. The objective of this work is to develop a heterogenous sensor data fusion approach capable of detecting multiscale flaws across different LPBF part geometries and build conditions. Accordingly, data from an infrared camera, spatter imaging camera, and optical powder bed imaging camera were acquired across separate builds with differing part geometries and orientations (Inconel 718). Spectral graph-based process signatures were extracted from this heterogeneous thermo-optical sensor data and used as inputs to simple machine learning models. The approach detected porosity, layer-level distortion, and geometry-related flaws with statistical fidelity exceeding 93% (F-score). [ABSTRACT FROM AUTHOR]

Preventing terrorist attacks at soft targets has become a priority for our society. The realization of sensor systems for automatic threat detection in crowded spaces, such as airports and metro stations, is challenged by the limited sensing coverage capability of the devices in place due to the variety of dangerous materials, to the scanning rate of the devices, and to the detection area covered. In this context, effectiveness of the physical configuration of the system based on the detectors used, the coordination of the sensor data collection, and the real time data analysis for threat identification and localization to enable timely reactions by the security guards are essential requirements for such integrated sensor-based applications. This paper describes a modular distributed architecture of a command-and-control software, which is independent from the specific detectors and where sensor data fusion is supported by two intelligent video systems. Furthermore, the system installation can be replicated at different locations of a public space. Person tracking and later re-identification in a separate area, and tracking hand-over between different video components, provide the command-and-control with localization information of threats to timely activate alarm management and support the activity of subsequent detectors. The architecture has been implemented for the NATO-funded DEXTER program and has been successfully tested in a big city trial at a metro station in Rome both when integrated with two real detectors of weapons and explosives and as a stand-alone system. The discussion focuses on the software functions of the command-and-control and on the flexibility and re-use of the system in wider settings. [ABSTRACT FROM AUTHOR]

Indoor fires pose significant threats in terms of casualties and economic losses globally. Thus, it is vital to accurately detect indoor fires at an early stage. To improve the accuracy of indoor fire detection for the resource-constrained embedded platform, an indoor fire detection method based on multi-sensor fusion and a lightweight convolutional neural network (CNN) is proposed. Firstly, the Savitzky–Golay (SG) filter is used to clean the three types of heterogeneous sensor data, then the cleaned sensor data are transformed by means of the Gramian Angular Field (GAF) method into matrices, which are finally integrated into a three-dimensional matrix. This preprocessing stage will preserve temporal dependency and enlarge the characteristics of time-series data. Therefore, we could reduce the number of blocks, channels and layers in the network, leading to a lightweight CNN for indoor fire detection. Furthermore, we use the Fire Dynamic Simulator (FDS) to simulate data for the training stage, enhancing the robustness of the network. The fire detection performance of the proposed method is verified through an experiment. It was found that the proposed method achieved an impressive accuracy of 99.1%, while the number of CNN parameters and the amount of computation is still small, which is more suitable for the resource-constrained embedded platform of an indoor fire detection system.

The rapid development of microsystems technology with the availability of various machine learning algorithms facilitates human activity recognition (HAR) and localization by low-cost and low-complexity systems in various applications related to industry 4.0, healthcare, ambient assisted living as well as tracking and navigation tasks. Previous work, which provided a spatiotemporal framework for HAR by fusing sensor data generated from an inertial measurement unit (IMU) with data obtained by an RGB photodiode for visible light sensing (VLS), already demonstrated promising results for real-time HAR and room identification. Based on these results, we extended the system by applying feature extraction methods of the time and frequency domain to improve considerably the correct determination of common human activities in industrial scenarios in combination with room localization. This increases the correct detection of activities to over 90% accuracy. Furthermore, it is demonstrated that this solution is applicable to real-world operating conditions in ambient light. [ABSTRACT FROM AUTHOR]

There is a lot of interference in underwater environment, which can affect underwater signal transmission. For this reason, an abnormal perception model of underwater signals based on multi-sensor data fusion is proposed. The underwater frequency hopping communication method is used to select the channel and construct the mathematical model of underwater communication signal. A signal modulation and recognition system for multi-sensor reception is established, and the multi-sensor data fusion model is obtained by calculating the support margin. The energy spectrum distribution is obtained by decomposition of abnormal features with wavelet detector in time domain. The deep learning algorithm is used to reorganize the signal with the minimum detection error, and the mathematical model of underwater signal anomaly perception is obtained. The experimental results show that the average accuracy of the method for locating abnormal underwater communication signals is 95.5%, the misrecognition rate of abnormal underwater signals is less than 2%, and the time-consuming for sensing abnormal signals of 420 MB underwater communication is less than 3 s. [ABSTRACT FROM AUTHOR]

Effective monitoring of the tool wear condition within a machining process can be very challenging. Depending on the sensors used, often only a part of the relevant wear information can be detected. In the case of milling processes data acquisition is made even more difficult by the fact that the process working point is inaccessible for sensor applications due to the physical tool, the machining process itself, the chipping and used cooling-lubricants. By using a variety of sensors and different measuring principles, sensor data fusion strategies can counteract this problem. An approach to this is the eigenface algorithm. This approach, a face recognition technique, is tested for its suitability on tool condition monitoring in milling processes by using multi-sensor process data. [ABSTRACT FROM AUTHOR]

Vehicle self-localization is one of the most important capabilities for automated driving. Current localization methods already provide accuracy in the centimeter range, so robustness becomes a key factor, especially in urban environments. There is no commonly used standard metric for the robustness of localization systems, but a set of different approaches. Here, we show a novel robustness score that combines different aspects of robustness and evaluate a graph-based localization method with the help of fault injections. In addition, we investigate the influence of semantic class information on robustness with a layered landmark model. By using the perturbation injections and our novel robustness score for test drives, system vulnerabilities or possible improvements are identified. Furthermore, we demonstrate that semantic class information allows early discarding of misclassified dynamic objects such as pedestrians, thus improving false-positive rates. This work provides a method for the robustness evaluation of landmark-based localization systems that are also capable of measuring the impact of semantic class information for vehicle self-localization.

This paper presents a leader–follower mobile robot control approach using onboard sensors. The follower robot is equipped with an Intel RealSense camera mounted on a rotating platform. Camera observations and ArUco markers are used to localize the robots to each other and relative to the workspace. The rotating platform allows the expansion of the perception range. As a result, the robot can use observations that are not within the camera’s field of view at the same time in the localization process. The decision-making process associated with the control of camera rotation is implemented using behavior trees. In addition, measurements from encoders and IMUs are used to improve the quality of localization. Data fusion is performed using the EKF filter and allows the user to determine the robot’s poses. A 3D-printed cuboidal tower is added to the leader robot with four ArUco markers located on its sides. Fiducial landmarks are placed on vertical surfaces in the workspace to improve the localization process. The experiments were performed to verify the effectiveness of the presented control algorithm. The robot operating system (ROS) was installed on both robots.

Ensemble learning systems (ELS) have been widely utilized for human activity recognition (HAR) with multiple homogeneous or heterogeneous sensors. However, traditional ensemble approaches for HAR cannot always work well due to insufficient accuracy and diversity of base classifiers, the absence of ensemble pruning, as well as the inefficiency of the fusion strategy. To overcome these problems, this paper proposes a novel selective ensemble approach with group decision-making (GDM) for decision-level fusion in HAR. As a result, the fusion process in the ELS is transformed into an abstract process that includes individual experts (base classifiers) making decisions with the GDM fusion strategy. Firstly, a set of diverse local base classifiers are constructed through the corresponding mechanism of the base classifier and the sensor. Secondly, the pruning methods and the number of selected base classifiers for the fusion phase are determined by considering the diversity among base classifiers and the accuracy of candidate classifiers. Two ensemble pruning methods are utilized: mixed diversity measure and complementarity measure. Thirdly, component decision information from the selected base classifiers is combined by using the GDM fusion strategy and the recognition results of the HAR approach can be obtained. Experimental results on two public activity recognition datasets (The OPPORTUNITY dataset; Daily and Sports Activity Dataset (DSAD)) suggest that the proposed GDM-based approach outperforms the well-known fusion techniques and other state-of-the-art approaches in the literature. [ABSTRACT FROM AUTHOR]

This study proposed a wearable inertial motion capture (Mocap) method for Parkinson’s lower limb motion reconstruction and analysis. To eliminate the error accumulation caused by sensor noise, the error state Kalman filter is used for attitude estimation. The inertial Mocap method proposed in this study is compared with the OptiTrack system to verify the motion tracking effect. The accuracy of different phase-detection methods is compared, and the highest recognition accuracy is 99.11% using the long short-term memory (LSTM) method. In addition, this study collected subjects in healthy control (HC), Parkinson’s disease (PD), and freezing of gait (FOG) groups for comparative analysis. The experimental results revealed differences in motion characteristics and kinematics between PD and healthy subjects. Thus, the quantitative analysis results can be used as an important reference for the clinical application of PD. [ABSTRACT FROM AUTHOR]

This paper presents a method to include unmodeled dynamics of load or a robot's end-effector into algorithms for collision detection or general understanding of a robot's operation context. The approach relies on the application of a previously developed modification of the Dynamic Time Warping algorithm, as well as a universally applicable algorithm for identifying kinematic parameters. The entire process can be applied to arbitrary robot configuration, and it does not require identification of dynamic parameters. The paper addresses the two main categories of contact tasks with unmodelled dynamics, which are determined based on whether the external contact force has a consistent profile in the end effector or base coordinate. Conclusions for representative examples analysed in the paper are applicable to tasks such as load manipulation, press bending, and crimping for the first type of forces and applications such as drilling, screwdriving, snap-fit, bolting, and riveting assembly for the latter category. The results presented in the paper are based on realistic testing with measurements obtained from an industrial robot. [ABSTRACT FROM AUTHOR]

Keywords: Measurement uncertainty determination; sensor data fusion; global reference system; Messunsicherheitsbestimmung; Sensordatenfusion; Globales Referenzsystem EN Measurement uncertainty determination sensor data fusion global reference system DE Messunsicherheitsbestimmung Sensordatenfusion Globales Referenzsystem 658 670 13 09/26/22 20221001 NES 221001 1 Current challenges in aeroplane assembly In aeroplane assembly, large components, such as fuselage sections and wings, are aligned and connected to each other [[12], p. 3]. In each time step HT ht , one prediction and one correction is carried out. The matrix HT ht can be determined via the above described assumed known relationships of the individual CS of the measuring systems and the assembly environment. For each entry of the pose HT ht , the sensor data listed in Table 2 are fused. [Extracted from the article]

In the context of intelligent components in industrial applications in the automotive, energy or construction sector, sensor monitoring is crucial for security issues and to avoid long and costly downtimes. This article discusses component-inherent thin-film sensors for this purpose, which, in contrast to conventional sensor technology, can be applied inseparably onto the component's surface via sputtering, so that a maximum of information about the component's condition can be generated, especially regarding deformation. This article examines whether the sensors can continue to generate reliable measurement data even after critical component loads have been applied. This extends their field of use concerning plastic deformation behavior. Therefore, any change in sensor properties is necessary for ongoing elastic strain measurements. These novel fundamentals are established for thin-film constantan strain gauges and platinum temperature sensors on steel substrates. In general, a k-factor decrease and an increase in the temperature coefficient of resistance with increasing plastic deformation could be observed until a sensor failure above 0.5% plastic deformation (constantan) occurred (1.3% for platinum). Knowing these values makes it possible to continue measuring elastic strains after critical load conditions on a machine component in terms of plastic deformation. Additionally, a method of sensor-data fusion for the clear determination of plastic deformation and temperature change is presented. [ABSTRACT FROM AUTHOR]

Increasingly complex automated driving functions, specifically those associated with free space detection (FSD), are delegated to convolutional neural networks (CNNs). If the dataset used to train the network lacks diversity, modality, or sufficient quantities, the driver policy that controls the vehicle may induce safety risks. Although most autonomous ground vehicles (AGVs) perform well in structured surroundings, the need for human intervention significantly rises when presented with unstructured niche environments. To this end, we developed an AGV for seamless indoor and outdoor navigation to collect realistic multimodal data streams. We demonstrate one application of the AGV when applied to a self-evolving FSD framework that leverages online active machine-learning (ML) paradigms and sensor data fusion. In essence, the self-evolving AGV queries image data against a reliable data stream, ultrasound, before fusing the sensor data to improve robustness. We compare the proposed framework to one of the most prominent free space segmentation methods, DeepLabV3+. DeepLabV3+ is a state-of-the-art semantic segmentation model composed of a CNN and an autodecoder. In consonance with the results, the proposed framework outperforms DeepLabV3+. The performance of the proposed framework is attributed to its ability to self-learn free space. This combination of online and active ML removes the need for large datasets typically required by a CNN. Moreover, this technique provides case-specific free space classifications based on the information gathered from the scenario at hand. [ABSTRACT FROM AUTHOR]