Search

With characters of low cost, portability, easy disposal, and high accuracy, as well as bulky reduced laboratory equipment, paper-based sensors are getting increasing attention for reliable indoor/outdoor onsite detection with nonexpert operation. They have become powerful analysis tools in trace detection with ultra-low detection limits and extremely high accuracy, resulting in their great popularity in medical detection, environmental inspection, and other applications. Herein, we summarize and generalize the recently reported paper-based sensors based on their application for mechanics, biomolecules, food safety, and environmental inspection. Based on the biological, physical, and chemical analytes-sensitive electrical or optical signals, extensive detections of a large number of factors such as humidity, pressure, nucleic acid, protein, sugar, biomarkers, metal ions, and organic/inorganic chemical substances have been reported via paper-based sensors. Challenges faced by the current paper-based sensors from the fundamental problems and practical applications are subsequently analyzed; thus, the future directions of paper-based sensors are specified for their rapid handheld testing.

We present a rotational terahertz imaging system for inline nondestructive testing (NDT) of press sleeves for the paper industry during fabrication. Press sleeves often consist of polyurethane (PU) which is deposited by rotational molding on metal barrels and its outer surface mechanically processed in several milling steps afterwards. Due to a stabilizing polyester fiber mesh inlay, small defects can form on the sleeve’s backside already during the initial molding, however, they cannot be visually inspected until the whole production processes is completed. We have developed a fast-scanning frequenc-modulated continuous wave (FMCW) terahertz imaging system, which can be integrated into the manufacturing process to yield high resolution images of the press sleeves and therefore can help to visualize hidden structural defects at an early stage of fabrication. This can save valuable time and resources during the production process. Our terahertz system can record images at 0.3 and 0.5 THz and we achieve data acquisition rates of at least 20 kHz, exploiting the fast rotational speed of the barrels during production to yield sub-millimeter image resolution. The potential of automated defect recognition by a simple machine learning approach for anomaly detection is also demonstrated and discussed.

Falls are unusual actions that cause a significant health risk among older people. The growing percentage of people of old age requires urgent development of fall detection and prevention systems. The emerging technology focuses on developing such systems to improve quality of life, especially for the elderly. A fall prevention system tries to predict and reduce the risk of falls. In contrast, a fall detection system observes the fall and generates a help notification to minimize the consequences of falls. A plethora of technical and review papers exist in the literature with a primary focus on fall detection. Similarly, several studies are relatively old, with a focus on wearables only, and use statistical and threshold-based approaches with a high false alarm rate. Therefore, this paper presents the latest research trends in fall detection and prevention systems using Machine Learning (ML) algorithms. It uses recent studies and analyzes datasets, age groups, ML algorithms, sensors, and location. Additionally, it provides a detailed discussion of the current trends of fall detection and prevention systems with possible future directions. This overview can help researchers understand the current systems and propose new methodologies by improving the highlighted issues. ispartof: SENSORS vol:21 issue:15 ispartof: location:Switzerland status: published

The increasing demand for smart vehicles with many sensing capabilities will escalate data traffic in vehicular networks. Meanwhile, available network resources are limited. The emergence of AI implementation in vehicular network resource allocation opens the opportunity to improve resource utilization to provide more reliable services. Accordingly, many resource allocation schemes with various machine learning algorithms have been proposed to dynamically manage and allocate network resources. This survey paper presents how machine learning is leveraged in the vehicular network resource allocation strategy. We focus our study on determining its role in the mechanism. First, we provide an analysis of how authors designed their scenarios to orchestrate the resource allocation strategy. Secondly, we classify the mechanisms based on the parameters they chose when designing the algorithms. Finally, we analyze the challenges in designing a resource allocation strategy in vehicular networks using machine learning. Therefore, a thorough understanding of how machine learning algorithms are utilized to offer a dynamic resource allocation in vehicular networks is provided in this study.

The field of human activity recognition (HAR) often utilizes wearable sensors and machine learning techniques in order to identify the actions of the subject. This paper considers the activity recognition of walking and running while using a support vector machine (SVM) that was trained on principal components derived from wearable sensor data. An ablation analysis is performed in order to select the subset of sensors that yield the highest classification accuracy. The paper also compares principal components across trials to inform the similarity of the trials. Five subjects were instructed to perform standing, walking, running, and sprinting on a self-paced treadmill, and the data were recorded while using surface electromyography sensors (sEMGs), inertial measurement units (IMUs), and force plates. When all of the sensors were included, the SVM had over 90% classification accuracy using only the first three principal components of the data with the classes of stand, walk, and run/sprint (combined run and sprint class). It was found that sensors that were placed only on the lower leg produce higher accuracies than sensors placed on the upper leg. There was a small decrease in accuracy when the force plates are ablated, but the difference may not be operationally relevant. Using only accelerometers without sEMGs was shown to decrease the accuracy of the SVM.

Outdoor vision sensing systems often struggle with poor weather conditions, such as snow and rain, which poses a great challenge to existing video desnowing and deraining methods. In this paper, we propose a novel video desnowing and deraining model that utilizes the salience information of moving objects to address this problem. First, we remove the snow and rain from the video by low-rank tensor decomposition, which makes full use of the spatial location information and the correlation between the three channels of the color video. Second, because existing algorithms often regard sparse snowflakes and rain streaks as moving objects, this paper injects salience information into moving object detection, which reduces the false alarms and missed alarms of moving objects. At the same time, feature point matching is used to mine the redundant information of moving objects in continuous frames, and a dual adaptive minimum filtering algorithm in the spatiotemporal domain is proposed by us to remove snow and rain in front of moving objects. Both qualitative and quantitative experimental results show that the proposed algorithm is more competitive than other state-of-the-art snow and rain removal methods.

Existing wearable systems that use G-sensors to identify daily activities have been widely applied for medical, sports and military applications, while body temperature as an obvious physical characteristic that has rarely been considered in the system design and relative applications of HAR. In the context of the normalization of COVID-19, the prevention and control of the epidemic has become a top priority. Temperature monitoring plays an important role in the preliminary screening of the population for fever. Therefore, this paper proposes a wearable device embedded with inertial and temperature sensors that is used to apply human behavior recognition (HAR) to body surface temperature detection for body temperature monitoring and adjustment by evaluating recognition algorithms. The sensing system consists of an STM 32-based microcontroller, a 6-axis (accelerometer and gyroscope) sensor, and a temperature sensor to capture the original data from 10 individual participants under 4 different daily activity scenarios. Then, the collected raw data are pre-processed by signal standardization, data stacking and resampling. For HAR, several machine learning (ML) and deep learning (DL) algorithms are implemented to classify the activities. To compare the performance of different classifiers on the seven-dimensional dataset with temperature sensing signals, evaluation metrics and the algorithm running time are considered, and random forest (RF) is found to be the best-performing classifier with 88.78% recognition accuracy, which is higher than the case of the absence of temperature data (&lt, 78%). In addition, the experimental results show that participants’ body surface temperature in dynamic activities was lower compared to sitting, which can be associated with the possible missing fever population due to temperature deviations in COVID-19 prevention. According to different individual activities, epidemic prevention workers are supposed to infer the corresponding standard normal body temperature of a patient by referring to the specific values of the mean expectation and variance in the normal distribution curve provided in this paper.

This paper presents a case study of continuous productivity improvement of an automotive parts production line using Internet of Everything (IoE) data for fault monitoring. Continuous productivity improvement denotes an iterative process of analyzing and updating the production line configuration for productivity improvement based on measured data. Analysis for continuous improvement of a production system requires a set of data (machine uptime, downtime, cycle-time) that are not typically monitored by a conventional fault monitoring system. Although productivity improvement is a critical aspect for a manufacturing site, not many production systems are equipped with a dedicated data recording system towards continuous improvement. In this paper, we study the problem of how to derive the dataset required for continuous improvement from the measurement by a conventional fault monitoring system. In particular, we provide a case study of an automotive parts production line. Based on the data measured by the existing fault monitoring system, we model the production system and derive the dataset required for continuous improvement. Our approach provides the expected amount of improvement to operation managers in a numerical manner to help them make a decision on whether they should modify the line configuration or not.

The development of the smart grid requires the distribution switch to not be limited to the original breaking function. More functional requirements lead to more complex switch structures, especially the intelligent processing unit on the secondary side. A technology called primary and secondary integration optimizes the structure of the switch, which greatly increases the intelligence level of the switch, but also has disadvantages. The secondary intelligent unit is arranged close to the primary high-voltage electromagnetic environment, and the distribution switch is prone to failure due to electromagnetic interference. In order to explore the influence of electromagnetic interference on it, a transient electromagnetic interference simulation test platform was built for a 10 kV intelligent distribution switch based on the principle of spherical gap arc discharge, and the interference signal of the intelligent distribution switch was measured; the law of the spatial magnetic field near the electronic transformer is mainly studied in this paper. The shielding effectiveness of the distribution terminal of the switch was analyzed, and the interference of the power line of the sensor merging unit circuit board was calculated. The results show that the electronic transformer may have serious faults under continuous strong transient electromagnetic interference. The electromagnetic transient simulation test system studied in this paper can evaluate the anti strong electromagnetic interference ability of the electronic transformer.

Effective communication in vehicular networks depends on the scheduling of wireless channel resources. There are two types of channel resource scheduling in Release 14 of the 3GPP, i.e., (1) controlled by eNodeB and (2) a distributed scheduling carried out by every vehicle, known as Autonomous Resource Selection (ARS). The most suitable resource scheduling for vehicle safety applications is the ARS mechanism. ARS includes (a) counter selection (i.e., specifying the number of subsequent transmissions) and (b) resource reselection (specifying the reuse of the same resource after counter expiry). ARS is a decentralized approach for resource selection. Therefore, resource collisions can occur during the initial selection, where multiple vehicles might select the same resource, hence resulting in packet loss. ARS is not adaptive towards vehicle density and employs a uniform random selection probability approach for counter selection and reselection. As a result, it can prevent some vehicles from transmitting in a congested vehicular network. To this end, the paper presents Truly Autonomous Resource Selection (TARS) for vehicular networks. TARS considers resource allocation as a problem of locally detecting the selected resources at neighbor vehicles to avoid resource collisions. The paper also models the behavior of counter selection and resource block reselection on resource collisions using the Discrete Time Markov Chain (DTMC). Observation of the model is used to propose a fair policy of counter selection and resource reselection in ARS. The simulation of the proposed TARS mechanism showed better performance in terms of resource collision probability and the packet delivery ratio when compared with the LTE Mode 4 standard and with a competing approach proposed by Jianhua He et al.

Global Positioning System (GPS) is a global navigation satellite system and the most common satellite system used in navigation and tracking devices. The phenomenon of week number rollover happened recently—a year ago—due to a design limitation in the week number variable that counting weeks which causes vast losses. As many fleet management systems depend on GPS raw data, such systems stopped working due to inaccurate data provided by GPS receivers. In this paper, we propose a technical and mathematical analysis for the GPS week number rollover phenomenon and suggest a solution to avoid the resulting damage to other subsystems that depend on the GPS device’s raw data. In addition, this paper seeks to provide precautionary measures to deal with the problem proactively. The Open Systems Interconnection model (OSI) and transport layer level solution that has been suggested depends on a TCP packet reforming tool that re-formats the value of the week number according to a mathematical model based on a timestamp complement. At the level of the database, a solution is also suggested which uses triggers. A hardware-level solution is suggested by applying a timestamp complement over the GPS internal controller. Complete testing is applied for all suggested solutions using actual data provided by Traklink—a leading company in navigation and fleet management solutions. After testing, it is evident that the transport layer level solution was the most effective in terms of speed, efficiency, accuracy, cost, and complexity. Applying a transport layer level complement mathematical model can fix the consequences of GPS week number rollover and provide stability to all subsystems that used GPS data from infected devices.

Recognizing Activities of Daily Living (ADL) or detecting falls in domestic environments require monitoring the movements and positions of a person. Several approaches use wearable devices or cameras, especially for fall detection, but they are considered intrusive by many users. To support such activities in an unobtrusive way, ambient-based solutions are available (e.g., based on PIRs, contact sensors, etc.). In this paper, we focus on the problem of sitting detection exploiting only unobtrusive sensors. In fact, sitting detection can be useful to understand the position of the user in many activities of the daily routines. While identifying sitting/lying on a sofa or bed is reasonably simple with pressure sensors, detecting whether a person is sitting on a chair is an open problem due to the natural chair position volatility. This paper proposes a reliable, not invasive and energetically sustainable system that can be used on chairs already present in the home. In particular, the proposed solution fuses the data of an accelerometer and a capacitive coupling sensor to understand if a person is sitting or not, discriminating the case of objects left on the chair. The results obtained in a real environment setting show an accuracy of 98.6% and a precision of 95%.

As an alternative approach, viseme-based lipreading systems have demonstrated promising performance results in decoding videos of people uttering entire sentences. However, the overall performance of such systems has been significantly affected by the efficiency of the conversion of visemes to words during the lipreading process. As shown in the literature, the issue has become a bottleneck of such systems where the system’s performance can decrease dramatically from a high classification accuracy of visemes (e.g., over 90%) to a comparatively very low classification accuracy of words (e.g., only just over 60%). The underlying cause of this phenomenon is that roughly half of the words in the English language are homophemes, i.e., a set of visemes can map to multiple words, e.g., “time” and “some”. In this paper, aiming to tackle this issue, a deep learning network model with an Attention based Gated Recurrent Unit is proposed for efficient viseme-to-word conversion and compared against three other approaches. The proposed approach features strong robustness, high efficiency, and short execution time. The approach has been verified with analysis and practical experiments of predicting sentences from benchmark LRS2 and LRS3 datasets. The main contributions of the paper are as follows: (1) A model is developed, which is effective in converting visemes to words, discriminating between homopheme words, and is robust to incorrectly classified visemes; (2) the model proposed uses a few parameters and, therefore, little overhead and time are required to train and execute; and (3) an improved performance in predicting spoken sentences from the LRS2 dataset with an attained word accuracy rate of 79.6%—an improvement of 15.0% compared with the state-of-the-art approaches.

This paper presents the empirical evaluation of the path-tracking accuracy of a three-wheeled omnidirectional mobile robot that is able to move in any direction while simultaneously changing its orientation. The mobile robot assessed in this paper includes a precise onboard LIDAR for obstacle avoidance, self-location and map creation, path-planning and path-tracking. This mobile robot has been used to develop several assistive services, but the accuracy of its path-tracking system has not been specifically evaluated until now. To this end, this paper describes the kinematics and path-planning procedure implemented in the mobile robot and empirically evaluates the accuracy of its path-tracking system that corrects the trajectory. In this paper, the information gathered by the LIDAR is registered to obtain the ground truth trajectory of the mobile robot in order to estimate the path-tracking accuracy of each experiment conducted. Circular and eight-shaped trajectories were assessed with different translational velocities. In general, the accuracy obtained in circular trajectories is within a short range, but the accuracy obtained in eight-shaped trajectories worsens as the velocity increases. In the case of the mobile robot moving at its nominal translational velocity, 0.3 m/s, the root mean square (RMS) displacement error was 0.032 m for the circular trajectory and 0.039 m for the eight-shaped trajectory, the absolute maximum displacement errors were 0.077 m and 0.088 m, with RMS errors in the angular orientation of 6.27° and 7.76°, respectively. Moreover, the external visual perception generated by these error levels is that the trajectory of the mobile robot is smooth, with a constant velocity and without perceiving trajectory corrections.

This paper considers the use of a post metadata-based approach to identifying intentionally deceptive online content. It presents the use of an inherently explainable artificial intelligence technique, which utilizes machine learning to train an expert system, for this purpose. It considers the role of three factors (textual context, speaker background, and emotion) in fake news detection analysis and evaluates the efficacy of using key factors, but not the inherently subjective processing of post text itself, to identify deceptive online content. This paper presents initial work on a potential deceptive content detection tool and also, through the networks that it presents for this purpose, considers the interrelationships of factors that can be used to determine whether a post is deceptive content or not and their comparative importance.

With the expansion of the Internet of Things (IoT), security incidents about exploiting vulnerabilities in IoT devices have become prominent. However, due to the characteristics of IoT devices such as low power and low performance, it is difficult to apply existing security solutions to IoT devices. As a result, IoT devices have easily become targets for cyber attackers, and malware attacks on IoT devices are increasing every year. The most representative is the Mirai malware that caused distributed denial of service (DDoS) attacks by creating a massive IoT botnet. Moreover, Mirai malware has been released on the Internet, resulting in increasing variants and new malicious codes. One of the ways to mitigate distributed denial of service attacks is to render the creation of massive IoT botnets difficult by preventing the spread of malicious code. For IoT infrastructure security, security solutions are being studied to analyze network packets going in and out of IoT infrastructure to detect threats, and to prevent the spread of threats within IoT infrastructure by dynamically controlling network access to maliciously used IoT devices, network equipment, and IoT services. However, there is a great risk to apply unverified security solutions to real-world environments. In this paper, we propose a malware simulation tool that scans vulnerable IoT devices assigned a private IP address, and spreads malicious code within IoT infrastructure by injecting malicious code download command into vulnerable devices. The malware simulation tool proposed in this paper can be used to verify the functionality of network threat detection and prevention solutions.

As is well-known, defects precisely affect the lives and functions of the machines in which they occur, and even cause potentially catastrophic casualties. Therefore, quality assessment before mounting is an indispensable requirement for factories. Apart from the recognition accuracy, current networks suffer from excessive computing complexity, making it of great difficulty to deploy in the manufacturing process. To address these issues, this paper introduces binary networks into the area of surface defect detection for the first time, for the reason that binary networks prohibitively constrain weight and activation to +1 and −1. The proposed Bi-ShuffleNet and U-BiNet utilize binary convolution layers and activations in low bitwidth, in order to reach comparable performances while incurring much less computational cost. Extensive experiments are conducted on real-life NEU and Magnetic Tile datasets, revealing the least OPs required and little accuracy decline. When classifying the defects, Bi-ShuffleNet yields comparable results to counterpart networks, with at least 2× inference complexity reduction. Defect segmentation results indicate similar observations. Some network design rules in defect detection and binary networks are also summarized in this paper.

This paper investigates damage identification metrics and their performance using a cantilever beam with a piezoelectric harvester for Structural Health Monitoring. In order to do this, the vibrations of three different beam structures are monitored in a controlled manner via two piezoelectric energy harvesters (PEH) located in two different positions. One of the beams is an undamaged structure recognized as reference structure, while the other two are beam structures with simulated damage in form of drilling holes. Subsequently, five different damage identification metrics for detecting damage localization and extent are investigated in this paper. Overall, each computational model has been designed on the basis of the modified First Order Shear Theory (FOST), considering an MFC element consisting homogenized materials in the piezoelectric fiber layer. Frequency response functions are established and five damage metrics are assessed, three of which are relevant for damage localization and the other two for damage extent. Experiments carried out on the lab stand for damage structure with control damage by using a modal hammer allowed to verify numerical results and values of particular damage metrics. In the effect, it is expected that the proposed method will be relevant for a wide range of application sectors, as well as useful for the evolving composite industry.

A drone-borne microwave radiometer requires a high sampling frequency and a continuous acquisition capability to detect and mitigate radio frequency interference (RFI), but existing methods cannot store such a large amount of data. In this paper, the dual polling write method (DPSM) for secure digital cards triggered by a timer under a multitask framework based on STM32 MCU is proposed to meet the requirements of continuous data storage. The card programming step was changed from a query waiting structure to a polling query flag bit structure, and time-sharing processing and parallel processing were used to simulate multithreading. The experimental results were as follows: (1) the time consumption of the whole storage procedure was reduced from 4000 microseconds to 200–400 microseconds, (2) the time consumption of the card programming step was reduced from 3000 microseconds in the first block and 1000 microseconds in the second and subsequent blocks to 17–174 microseconds and 18–71 microseconds, respectively, compared with the existing method, (3) the delay in the whole sampling cycle was reduced from 3942 microseconds to 0 microseconds. The results of this paper can meet the data storage requirements of a drone-borne microwave radiometer and be applied to the high-speed storage of other devices.

Human-robot interaction has received a lot of attention as collaborative robots became widely utilized in many industrial fields. Among techniques for human-robot interaction, collision identification is an indispensable element in collaborative robots to prevent fatal accidents. This paper proposes a deep learning method for identifying external collisions in 6-DoF articulated robots. The proposed method expands the idea of CollisionNet, which was previously proposed for collision detection, to identify the locations of external forces. The key contribution of this paper is uncertainty-aware knowledge distillation for improving the accuracy of a deep neural network. Sample-level uncertainties are estimated from a teacher network, and larger penalties are imposed for uncertain samples during the training of a student network. Experiments demonstrate that the proposed method is effective for improving the performance of collision identification.

There are a dozen messaging protocols proposed for IoT systems. Choosing one for a new design is complicated, and a non-optimal selection can result in slower development and higher design costs. This paper aims to help select appropriate protocols, considering IoT applications’ specificity and communication requirements. We have identified the protocol features that are significant for the design and operation of IoT systems. This paper gives a substantial comparison of the protocols using the features and is based on a thorough analysis of the protocol specifications. The results contain an assessment of the suitability of the protocols for the defined types of IoT devices and the identified communication purposes. We conclude the comparison with some recommendations of the protocol selection and usage.

As a typical application of sensor networks, there exist many information security problems in smart parks, such as confusion of personnel access, lack of security management, disorderly data flow, insufficient collection of audit evidence, and so on. Aiming at the scenario of personnel and equipment moving in different areas of smart parks, the paper proposes a joint authorization and dynamic access control mechanism, which can provide unified identity management services, access control services, and policy management services, and effectively solve the problem of multi-authorization in user identity and authority management. The license negotiation interaction protocol is designed to prevent common network attack threats in the process of identity authentication and authority management. In order to realize the tamper-proof storage of personnel and equipment movement trajectory, the paper also designs a movement trajectory traceability protocol based on a Merkle tree, which solves the problems of internal personnel malicious attack, trusted third-party dependency bottleneck, high overheads of tracking algorithms, and so on. The experimental results show that compared with the current security control mechanisms for sensor networks, the joint authorization, and dynamic access control mechanism can support multi-party authorization and traceability, while the overhead it generates in initialization, encryption, decryption, and key generation steps are basically the same as other mechanisms do.

Autonomous Vehicles (AVs) have the potential to solve many traffic problems, such as accidents, congestion and pollution. However, there are still challenges to overcome, for instance, AVs need to accurately perceive their environment to safely navigate in busy urban scenarios. The aim of this paper is to review recent articles on computer vision techniques that can be used to build an AV perception system. AV perception systems need to accurately detect non-static objects and predict their behaviour, as well as to detect static objects and recognise the information they are providing. This paper, in particular, focuses on the computer vision techniques used to detect pedestrians and vehicles. There have been many papers and reviews on pedestrians and vehicles detection so far. However, most of the past papers only reviewed pedestrian or vehicle detection separately. This review aims to present an overview of the AV systems in general, and then review and investigate several detection computer vision techniques for pedestrians and vehicles. The review concludes that both traditional and Deep Learning (DL) techniques have been used for pedestrian and vehicle detection; however, DL techniques have shown the best results. Although good detection results have been achieved for pedestrians and vehicles, the current algorithms still struggle to detect small, occluded, and truncated objects. In addition, there is limited research on how to improve detection performance in difficult light and weather conditions. Most of the algorithms have been tested on well-recognised datasets such as Caltech and KITTI; however, these datasets have their own limitations. Therefore, this paper recommends that future works should be implemented on more new challenging datasets, such as PIE and BDD100K. EPSRC DTP PhD studentship

With the automotive industry moving towards automated driving, sensing is increasingly important in enabling technology. The virtual sensors allow data fusion from various vehicle sensors and provide a prediction for measurement that is hard or too expensive to measure in another way or in the case of demand on continuous detection. In this paper, virtual sensing is discussed for the case of vehicle suspension control, where information about the relative velocity of the unsprung mass for each vehicle corner is required. The corresponding goal can be identified as a regression task with multi-input sequence input. The hypothesis is that the state-of-art method of Bidirectional Long–Short Term Memory (BiLSTM) can solve it. In this paper, a virtual sensor has been proposed and developed by training a neural network model. The simulations have been performed using an experimentally validated full vehicle model in IPG Carmaker. Simulations provided the reference data which were used for Neural Network (NN) training. The extensive dataset covering 26 scenarios has been used to obtain training, validation and testing data. The Bayesian Search was used to select the best neural network structure using root mean square error as a metric. The best network is made of 167 BiLSTM, 256 fully connected hidden units and 4 output units. Error histograms and spectral analysis of the predicted signal compared to the reference signal are presented. The results demonstrate the good applicability of neural network-based virtual sensors to estimate vehicle unsprung mass relative velocity.

The literature on coronaviruses counts more than 300,000 publications. Finding relevant papers concerning arbitrary queries is essential to discovery helpful knowledge. Current best information retrieval (IR) use deep learning approaches and need supervised training sets with labeled data, namely to know a priori the queries and their corresponding relevant papers. Creating such labeled datasets is time-expensive and requires prominent experts’ efforts, resources insufficiently available under a pandemic time pressure. We present a new self-supervised solution, called SUBLIMER, that does not require labels to learn to search on corpora of scientific papers for most relevant against arbitrary queries. SUBLIMER is a novel efficient IR engine trained on the unsupervised COVID-19 Open Research Dataset (CORD19), using deep metric learning. The core point of our self-supervised approach is that it uses no labels, but exploits the bibliography citations from papers to create a latent space where their spatial proximity is a metric of semantic similarity, for this reason, it can also be applied to other domains of papers corpora. SUBLIMER, despite is self-supervised, outperforms the Precision@5 (P@5) and Bpref of the state-of-the-art competitors on CORD19, which, differently from our approach, require both labeled datasets and a number of trainable parameters that is an order of magnitude higher than our.

As a complex task, robot sorting has become a research hotspot. In order to enable robots to perform simple, efficient, stable and accurate sorting operations for stacked multi-objects in unstructured scenes, a robot multi-object sorting system is built in this paper. Firstly, the training model of rotating target detection is constructed, and the placement state of five common objects in unstructured scenes is collected as the training set for training. The trained model is used to obtain the position, rotation angle and category of the target object. Then, the instance segmentation model is constructed, and the same data set is made, and the instance segmentation network model is trained. Then, the optimized Mask R-CNN instance segmentation network is used to segment the object surface pixels, and the upper surface point cloud is extracted to calculate the normal vector. Then, the angle obtained by the normal vector of the upper surface and the rotation target detection network is fused with the normal vector to obtain the attitude of the object. At the same time, the grasping order is calculated according to the average depth of the surface. Finally, after the obtained object posture, category and grasping sequence are fused, the performance of the rotating target detection network, the instance segmentation network and the robot sorting system are tested on the established experimental platform. Based on this system, this paper carried out an experiment on the success rate of object capture in a single network and an integrated network. The experimental results show that the multi-object sorting system based on deep learning proposed in this paper can sort stacked objects efficiently, accurately and stably in unstructured scenes.

This paper deals with the design of a DC motor speed control implemented by an embedded controller. The design is simple and brings some important changes to the traditional Ziegler–Nichols tuning. The design also includes a novel anti-windup implementation of the controller and an integrated noise-reduction filter design. The proposed tuning method considers all important aspects of the control, such as pre-processing of the measured signals and filtering (to attenuate the measurement noise), time delays of the process, modeling and identification of the process, and constraints on the control signal. Three important aspects of designing PI and PID controllers for processes with noisy output on Arduino-type embedded computers are considered. First, it deals with the integrated design of the input filter and the controller parameters, since both are interdependent. Secondly, the method of setting the controllers from step responses by Ziegler and Nichols is modified for the case of digital signal processing (without drawing the tangent), while it recommends the suitability of its modification in terms of the use of both integral and static models. Third, the most suitable anti-windup solution for the given controller structure is proposed. In summary, the paper shows that an appropriate design of the embedded controller can achieve excellent closed-loop performance even in a noisy process environment with limited control signals.

Four new algorithms (RFTA1, RFTA2, GFGF2A, and RFTA2GE) handling the event in wireless sensor and robot networks based on the Greedy-Face-Greedy (GFG) routing extended with auctions are proposed in this paper. In this paper we assume that all robots are mobile and after the event is found (reported by sensors), the goal is to allocate the task to the most suitable robot to act upon the event, using either distance or the robots' remaining energy as metrics. The proposed algorithms consist of two phases. The first phase of algorithms is based on face routing and we introduced the parameter called search radius (SR) at the end of this first phase. Routing is considered successful if the found robot is inside SR. After that, the second phase, based on auctions, is initiated by the robot found in SR trying to find a more suitable one. In the simulations, network lifetime and communication costs are measured and used for comparison. We compare our algorithms with two similar algorithms from the literature (k-SAAP and BFS) used for the task assignment. RFTA2 and RFTA2GE feature up to 7 times longer network lifetime with significant communication overhead reduction compared to k-SAAP and BFS. Among our algorithms, RFTA2GE features the best robot energy utilization., Comment: 21 p., submitted for journal publication

This paper studies the usage of orthogonal frequency division multiple access (OFDMA) for uplink transmissions in IEEE 802.11ax networks. OFDMA enables simultaneous multi-user transmissions in Wi-Fi, but its usage requires efficient resource allocation algorithms. These algorithms should be able to adapt to the changing channel conditions, including the frequency-selective fading. This paper presents an OFDMA resource allocation algorithm for channels with frequency-selective fading and proposes an approach to adapt the user transmission power and modulation and coding schemes to the varying channel conditions, which is efficient even in the case when the access point has outdated channel state information. The proposed scheduling algorithm and power allocation approach can double the goodput and halve the data transmission time in Wi-Fi networks even in dense deployments of access points.

In recent years, interest in home energy management systems (HEMS) has grown significantly, as well as the development of Voice Assistants that substantially increase home comfort. This paper presents a novel merging of HEMS with the Assistant paradigm. The combination of both concepts has allowed the creation of a high-performance and easy-to-manage expert system (ES). It has been developed in a framework that includes, on the one hand, the efficient energy management functionality boosted with an Internet of Things (IoT) platform, where artificial intelligence (AI) and big data treatment are blended, and on the other hand, an assistant that interacts both with the user and with the HEMS itself. The creation of this ES has made it possible to optimize consumption levels, improve security, efficiency, comfort, and user experience, as well as home security (presence simulation or security against intruders), automate processes, optimize resources, and provide relevant information to the user facilitating decision making, all based on a multi-objective optimization (MOP) problem model. This paper presents both the scheme and the results obtained, the synergies generated, and the conclusions that can be drawn after 24 months of operation.

Uncertainty in dense heterogeneous IoT sensor networks can be decreased by applying reputation-inspired algorithms, such as the EWMA (Exponentially Weighted Moving Average) algorithm, which is widely used in social networks. Despite its popularity, the eventual convergence of this algorithm for the purpose of IoT networks has not been widely studied, and results of simulations are often taken in lieu of the more rigorous proof. Therefore the question remains, whether under stable conditions, in realistic situations found in IoT networks, this algorithm indeed converges. This paper demonstrates proof of the eventual convergence of the EWMA algorithm. The proof consists of two steps: it models the sensor network as the UOG (Uniform Opinion Graph) that enables the analytical approach to the problem, and then offers the mathematical proof of eventual convergence, using formalizations identified in the previous step. The paper demonstrates that the EWMA algorithm converges under all realistic conditions.

This invited featured paper offers a Doctrinal Conception of sensing using Light (SuL) as an "umbrella" in which any sensing approach using Light Sciences and Technologies can be easily included. The key requirements of a sensing system will be quickly introduced by using a bottom-up methodology. Thanks to this, it will be possible to get a general conception of a sensor using Light techniques and know some related issues, such as its main constituted parts and types. The case in which smartness is conferred to the device is also considered. A quick "flight" over 10 significant cases using different principles, techniques, and technologies to detect diverse measurands in various sector applications is offered to illustrate this general concept. After reading this paper, any sensing approach using Light Sciences and Technologies may be easily included under the umbrella: sensing using Light or photonic sensors (PS). This work has been supported by Ministerio de Ciencia e Innovación and Agencia Estatal de Investigación (PID2019-107270RB-C21/AIE/10.13039/501100011033) and also TeDFeS Project (RTC-2017-6321-1) co-funded by European FEDER funds.

The proposed StegoFrameOrder (SFO) method enables the transmission of covert data in wireless computer networks exploiting non-deterministic algorithms of medium access (such as the distributed coordination function), especially in IEEE 802.11 networks. Such a covert channel enables the possibility of leaking crucial information outside secured network in a manner that is difficult to detect. The SFO method embeds hidden bits of information in the relative order of frames transmitted by wireless terminals operating on the same radio channel. The paper presents an idea of this covert channel, its implementation, and possible variants. The paper also discusses implementing the SFO method in a real environment and the experiments performed in the real-world scenario.

Many methods such as biomechanics and coaching have been proposed to help people learn a certain movement. There have been proposals for methods to discover characteristics of movement based on information obtained from videos and sensors. Especially in sports, it is expected that these methods can provide hints to improve movement skills. However, conventional methods focus on individual movements, and do not consider cases where external factors influence the movement, such as combat sports. In this paper, we propose a novel method called the Extraction for Successful Movement method (XSM method). Applying the method, this paper focuses on throwing techniques in judo to discover key factors that induce successful throwing from the postures right before initiating the throwing techniques. We define candidate factors by observing the video scenes where the throwing techniques are successfully performed. The method demonstrates the significance of the key factors according to the predominance of factors by χ2 test and residual analysis. Applying the XSM method to the dataset obtained from the videos of the Judo World Championships, we demonstrate the validity of the method with discussing the key factors related to the successful throwing techniques.

Forestry is a complex economic sector which is relying on resource and process monitoring data. Most of the forest operations such as planting and harvesting are supported by the use of tools and machines, and their monitoring has been traditionally done by the use of pen-and-paper time studies. Nevertheless, modern data collection and analysis methods involving different kinds of platforms and machine learning techniques have been studied lately with the aim of easing the data management process. By their outcomes, improvements are still needed to reach a close to 100% activity recognition, which may depend on several factors such as the type of monitored process and the characteristics of the signals used as inputs. In this paper, we test, thought a case study on mechanized pit-drilling operations, the potential of digital signal processing techniques combined with Artificial Neural Networks (ANNs) in improving the event-based classification accuracy in the time domain. Signal processing was implemented by the means of median filtering of triaxial accelerometer data (window sizes of 3, 5, and up to 21 observations collected at 1 Hz) while the ANNs were subjected to the regularization hyperparameter’s tunning. An acceleration signal processed by a median filter with a window size of 3 observations and fed into an ANN set to learn and generalize by a regularization parameter of α = 0.01 has been found to be the best strategy in improving the event-based classification accuracy (improvements of 1 to 8% in classification accuracy depending on the type of event in question). Improvement of classification accuracy by signal filtering and ANN tuning may depend largely on the type of monitored process and its outcomes in terms of event duration, therefore, other monitoring applications may need particular designs of signal processing and ANN tuning.

This paper describes the calibration method for calculating parameters (position and orientation) of planar reflectors reshaping LiDAR’s (light detection and ranging) field of view. The calibration method is based on the reflection equation used in the ICP (Iterative Closest Point) optimization. A novel calibration process as the multi-view data registration scheme is proposed, therefore, the poses of the measurement instrument and parameters of planar reflectors are calculated simultaneously. The final metric measurement is more accurate compared with parameters retrieved from the mechanical design. Therefore, it is evident that the calibration process is required for affordable solutions where the mechanical design can differ from the inaccurate assembly. It is shown that the accuracy is less than 20 cm for almost all measurements preserving long-range capabilities. The experiment is performed based on Livox Mid-40 LiDAR augmented with six planar reflectors. The ground-truth data were collected using Z + F IMAGER 5010 3D Terrestrial Laser Scanner. The calibration method is independent of mechanical design and does not require any fiducial markers on the mirrors. This work fulfils the gap between rotating and Solid-State LiDARs since the field of view can be reshaped by planar reflectors, and the proposed method can preserve the metric accuracy. Thus, such discussion concludes the findings. We prepared an open-source project and provided all the necessary data for reproducing the experiments. That includes: Complete open-source code, the mechanical design of reflector assembly and the dataset which was used in this paper.

In order to meet the assist requirements of extravehicular activity (EVA) for astronauts, such as moving outside the international space station (ISS) or performing on-orbit tasks by a single astronaut, this paper proposes an astronaut robotic limbs system (AstroLimbs) for extravehicular activities assistance. This system has two robotic limbs that can be fixed on the backpack of the astronaut. Each limb is composed of several basic module units with identical structure and function, which makes it modularized and reconfigurable. The robotic limbs can work as extra arms of the astronaut to assist them outside the space station cabin. In this paper, the robotic limbs are designed and developed. The reinforcement learning method is introduced to achieve autonomous motion planning capacity for the robot, which makes the robot intelligent enough to assist the astronaut in unstructured environment. In the meantime, the movement of the robot is also planned to make it move smoothly. The structure scene of the ISS for extravehicular activities is modeled in a simulation environment, which verified the effectiveness of the proposed method.

In this paper, we will introduce a method for observing microvascular waves (MVW) by extracting different images from the available images in the video taken with consumer cameras. Microvascular vasomotion is a dynamic phenomenon that can fluctuate over time for a variety of reasons and its sensing is used for variety of purposes. The special device, a side stream dark field camera (SDF camera) was developed in 2015 for the medical purpose to observe blood flow from above the epidermis. However, without using SDF cameras, smart signal processing can be combined with a consumer camera to analyze the global motion of microvascular vasomotion. MVW is a propagation pattern of microvascular vasomotions which reflects biological properties of vascular network. In addition, even without SDF cameras, MVW can be analyzed as a spatial and temporal pattern of microvascular vasomotion using a combination of advanced signal processing with consumer cameras. In this paper, we will demonstrate that such vascular movements and MVW can be observed using a consumer cameras. We also show a classification using it.

Future distribution grids will be subjected to fluctuations in voltages and power flows due to the presence of renewable sources with intermittent power generation. The advanced smart metering infrastructure (AMI) enables the distribution system operators (DSOs) to measure and analyze electrical quantities such as voltages, currents and power at each customer connection point. Various smart grid applications can make use of the AMI data either in offline or close to real-time mode to assess the grid voltage conditions and estimate losses in the lines/cables. The outputs of these applications can enable DSOs to take corrective action and make a proper plan for grid upgrades. In this paper, the process of development and deployment of applications for improving the observability of distributions grids is described, which consists of the novel deployment framework that encompasses the proposition of data collection, communication to the servers, data storage, and data visualization. This paper discussed the development of two observability applications for grid monitoring and loss calculation, their validation in a laboratory setup, and their field deployment. A representative distribution grid in Denmark is chosen for the study using an OPAL-RT real-time simulator. The results of the experimental studies show that the proposed applications have high accuracy in estimating grid voltage magnitudes and active energy losses. Further, the field deployment of the applications prove that DSOs can gain insightful information about their grids and use them for planning purposes.

In recent years, the rapid development of Deep Learning (DL) has provided a new method for ship detection in Synthetic Aperture Radar (SAR) images. However, there are still four challenges in this task. (1) The ship targets in SAR images are very sparse. A large number of unnecessary anchor boxes may be generated on the feature map when using traditional anchor-based detection models, which could greatly increase the amount of computation and make it difficult to achieve real-time rapid detection. (2) The size of the ship targets in SAR images is relatively small. Most of the detection methods have poor performance on small ships in large scenes. (3) The terrestrial background in SAR images is very complicated. Ship targets are susceptible to interference from complex backgrounds, and there are serious false detections and missed detections. (4) The ship targets in SAR images are characterized by a large aspect ratio, arbitrary direction and dense arrangement. Traditional horizontal box detection can cause non-target areas to interfere with the extraction of ship features, and it is difficult to accurately express the length, width and axial information of ship targets. To solve these problems, we propose an effective lightweight anchor-free detector called R-Centernet+ in the paper. Its features are as follows: the Convolutional Block Attention Module (CBAM) is introduced to the backbone network to improve the focusing ability on small ships, the Foreground Enhance Module (FEM) is used to introduce foreground information to reduce the interference of the complex background, the detection head that can output the ship angle map is designed to realize the rotation detection of ship targets. To verify the validity of the proposed model in this paper, experiments are performed on two public SAR image datasets, i.e., SAR Ship Detection Dataset (SSDD) and AIR-SARShip. The results show that the proposed R-Centernet+ detector can detect both inshore and offshore ships with higher accuracy than traditional models with an average precision of 95.11% on SSDD and 84.89% on AIR-SARShip, and the detection speed is quite fast with 33 frames per second.

In this paper, Computer Vision (CV) sensing technology based on Convolutional Neural Network (CNN) is introduced to process topographic maps for predicting wireless signal propagation models, which are applied in the field of forestry security monitoring. In this way, the terrain-related radio propagation characteristic including diffraction loss and shadow fading correlation distance can be predicted or extracted accurately and efficiently. Two data sets are generated for the two prediction tasks, respectively, and are used to train the CNN. To enhance the efficiency for the CNN to predict diffraction losses, multiple output values for different locations on the map are obtained in parallel by the CNN to greatly boost the calculation speed. The proposed scheme achieved a good performance in terms of prediction accuracy and efficiency. For the diffraction loss prediction task, 50% of the normalized prediction error was less than 0.518%, and 95% of the normalized prediction error was less than 8.238%. For the correlation distance extraction task, 50% of the normalized prediction error was less than 1.747%, and 95% of the normalized prediction error was less than 6.423%. Moreover, diffraction losses at 100 positions were predicted simultaneously in one run of CNN under the settings in this paper, for which the processing time of one map is about 6.28 ms, and the average processing time of one location point can be as low as 62.8 us. This paper shows that our proposed CV sensing technology is more efficient in processing geographic information in the target area. Combining a convolutional neural network to realize the close coupling of a prediction model and geographic information, it improves the efficiency and accuracy of prediction.

Currently, simultaneous localization and mapping (SLAM) is one of the main research topics in the robotics field. Visual-inertia SLAM, which consists of a camera and an inertial measurement unit (IMU), can significantly improve robustness and enable scale weak-visibility, whereas monocular visual SLAM is scale-invisible. For ground mobile robots, the introduction of a wheel speed sensor can solve the scale weak-visibility problem and improve robustness under abnormal conditions. In this paper, a multi-sensor fusion SLAM algorithm using monocular vision, inertia, and wheel speed measurements is proposed. The sensor measurements are combined in a tightly coupled manner, and a nonlinear optimization method is used to maximize the posterior probability to solve the optimal state estimation. Loop detection and back-end optimization are added to help reduce or even eliminate the cumulative error of the estimated poses, thus ensuring global consistency of the trajectory and map. The outstanding contribution of this paper is that the wheel odometer pre-integration algorithm, which combines the chassis speed and IMU angular speed, can avoid the repeated integration caused by linearization point changes during iterative optimization, state initialization based on the wheel odometer and IMU enables a quick and reliable calculation of the initial state values required by the state estimator in both stationary and moving states. Comparative experiments were conducted in room-scale scenes, building scale scenes, and visual loss scenarios. The results showed that the proposed algorithm is highly accurate—2.2 m of cumulative error after moving 812 m (0.28%, loopback optimization disabled)—robust, and has an effective localization capability even in the event of sensor loss, including visual loss. The accuracy and robustness of the proposed method are superior to those of monocular visual inertia SLAM and traditional wheel odometers.

Recent advancements in information and communication technologies (ICT) have im‐proved the power grid, leading to what is known as the smart grid, which, as part of a critical economic and social infrastructure, is vulnerable to security threats from the use of ICT and new emerging vulner‐abilities and privacy issues. Access control is a fundamental element of a security infrastructure, and se‐curity is based on the principles of less privilege, zero‐trust, and segregation of duties. This work ad‐dresses how access control can be applied without disrupting the power grid’s functioning while also properly maintaining the security, scalability, and interoperability of the smart grid. The authentication in the platform presumes digital certificates using a web of trust. This paper presents the findings of the SealedGRID project, and the steps taken for implementing Attribute‐based access control policies specif‐ically customized to the smart grid. The outcome is to develop a novel, hierarchical architecture com‐posed of different licensing entities that manages access to resources within the network infrastructure. They are based on well‐drawn policy rules and the security side of these resources is placed through a context awareness module. Together with this technology, the IoT is used with Big Data (facilitating easy handling of large databases). Another goal of this paper is to present implementation and evaluations details of a secure and scalable security platform for the smart grid.

A high power-conversion-efficiency voltage boost converter with MPPT for wireless sensor nodes (WSNs) is proposed in this paper. Since tiny wireless sensor nodes are all over complex environments, an efficient power management system (PMS) must be equipped to achieve long-term self-power supply and maintain regular operation. It is common to use Photovoltaic cells (PV) to harvest sunlight in the environment. However, most existing interface boost integrated circuits for the PV cell have low efficiency. This paper presents a voltage boost converter (VBC) with high power conversion efficiency (PCE) for WSNs. The integrated circuit (IC) designed in this paper includes a novel four-phase high-efficiency charge pump module, an ultra-low-power perturbation observation (P&O) MPPT control circuit module, a feedback control module, a nano-ampere current reference, etc. Manufactured in a standard 0.35 um complementary metal-oxide-semiconductor (CMOS) technology, the chip area is 3.15 mm × 2.43 mm. Test results demonstrate that when the output voltage of the PV cell is more than 0.5 V, VBC can improve the voltage to 3Vin, and the calculated voltage conversion efficiency can reach 99.4%. P&O MPPT algorithm makes output power improving 8.53%. Furthermore, when the output load current is 297uA, the output PCE achieves 85.1%.

The development of low-cost mass sensors is of unique interest for the scientific community due to the wide range of fields requiring these kind of devices. In this paper, a full inkjet-printed mass sensor is proposed. The device is based on a PolyEthylene Terephthalate (PET) cantilever beam (operating in its first natural frequency) where a strain-sensor and a planar coil have been realized by a low-cost InkJet Printing technology to implement the sensing and actuation strategies, respectively. The frequency readout strategy of the sensor presents several advantages, such as the intrinsic robustness against instabilities of the strain sensor, the residual stress of the cantilever beam, the target mass material, and the distance between the permanent magnet and the actuation coil (which changes as a function of the target mass values). However, the frictionless actuation mode represents another shortcoming of the sensor. The paper describes the sensor design, realization, and characterization while investigating its expected behavior by exploiting dedicate models. The working span of the device is 0–0.36 g while its resolution is in the order of 0.001 g, thus addressing a wide range of potential applications requiring very accurate mass measurements within a narrow operating range.

This paper presents a non-random weight initialization method in convolutional layers of neural networks examined with the Fast Gradient Sign Method (FSGM) attack. This paper's focus is convolutional layers, and are the layers that have been responsible for better than human performance in image categorization. The proposed method induces earlier learning through the use of striped forms, and as such has less unlearning of the existing random number speckled methods, consistent with the intuitions of Hubel and Wiesel. The proposed method provides a higher performing accuracy in a single epoch, with improvements of between 3-5% in a well known benchmark model, of which the first epoch is the most relevant as it is the epoch after initialization. The proposed method is also repeatable and deterministic, as a desirable quality for safety critical applications in image classification within sensors. That method is robust to Glorot/Xavier and He initialization limits as well. The proposed non-random initialization was examined under adversarial perturbation attack through the FGSM approach with transferred learning, as a technique to measure the affect in transferred learning with controlled distortions, and finds that the proposed method is less compromised to the original validation dataset, with higher distorted datasets.

In intelligent vehicles, extrinsic camera calibration is preferable to be conducted on a regular basis to deal with unpredictable mechanical changes or variations on weight load distribution. Specifically, high-precision extrinsic parameters between the camera coordinate and the world coordinate are essential to implement high-level functions in intelligent vehicles such as distance estimation and lane departure warning. However, conventional calibration methods, which solve a Perspective-n-Point problem, require laborious work to measure the positions of 3D points in the world coordinate. To reduce this inconvenience, this paper proposes an automatic camera calibration method based on 3D reconstruction. The main contribution of this paper is a novel reconstruction method to recover 3D points on planes perpendicular to the ground. The proposed method jointly optimizes reprojection errors of image features projected from multiple planar surfaces, and finally, it significantly reduces errors in camera extrinsic parameters. Experiments were conducted in synthetic simulation and real calibration environments to demonstrate the effectiveness of the proposed method.

Video applications have become one of the major services in the engineering field, which are implemented by server–client systems connected via the Internet, broadcasting services for mobile devices such as smartphones and surveillance cameras for security. Recently, the majority of video encoding mechanisms to reduce the data rate are mainly lossy compression methods such as the MPEG format. However, when we consider special needs for high-speed communication such as display applications and object detection ones with high accuracy from the video stream, we need to address the encoding mechanism without any loss of pixel information, called visually lossless compression. This paper focuses on the Adaptive Differential Pulse Code Modulation (ADPCM) that encodes a data stream into a constant bit length per data element. However, the conventional ADPCM does not have any mechanism to control dynamically the encoding bit length. We propose a novel ADPCM that provides a mechanism with a variable bit-length control, called ADPCM-VBL, for the encoding/decoding mechanism. Furthermore, since we expect that the encoded data from ADPCM maintains low entropy, we expect to reduce the amount of data by applying a lossless data compression. Applying ADPCM-VBL and a lossless data compression, this paper proposes a video transfer system that controls throughput autonomously in the communication data path. Through evaluations focusing on the aspects of the encoding performance and the image quality, we confirm that the proposed mechanisms effectively work on the applications that needs visually lossless compression by encoding video stream in low latency.

This paper addresses the problem of unsupervised video summarization. Video summarization helps people browse large-scale videos easily with a summary from the selected frames of the video. In this paper, we propose an unsupervised video summarization method with piecewise linear interpolation (Interp-SUM). Our method aims to improve summarization performance and generate a natural sequence of keyframes with predicting importance scores of each frame utilizing the interpolation method. To train the video summarization network, we exploit a reinforcement learning-based framework with an explicit reward function. We employ the objective function of the exploring under-appreciated reward method for training efficiently. In addition, we present a modified reconstruction loss to promote the representativeness of the summary. We evaluate the proposed method on two datasets, SumMe and TVSum. The experimental result showed that Interp-SUM generates the most natural sequence of summary frames than any other the state-of-the-art methods. In addition, Interp-SUM still showed comparable performance with the state-of-art research on unsupervised video summarization methods, which is shown and analyzed in the experiments of this paper.

Recent progress in deep learning has led to accurate and efficient generic object detection networks. Training of highly reliable models depends on large datasets with highly textured and rich images. However, in real-world scenarios, the performance of the generic object detection system decreases when (i) occlusions hide the objects, (ii) objects are present in low-light images, or (iii) they are merged with background information. In this paper, we refer to all these situations as challenging environments. With the recent rapid development in generic object detection algorithms, notable progress has been observed in the field of deep learning-based object detection in challenging environments. However, there is no consolidated reference to cover the state of the art in this domain. To the best of our knowledge, this paper presents the first comprehensive overview, covering recent approaches that have tackled the problem of object detection in challenging environments. Furthermore, we present a quantitative and qualitative performance analysis of these approaches and discuss the currently available challenging datasets. Moreover, this paper investigates the performance of current state-of-the-art generic object detection algorithms by benchmarking results on the three well-known challenging datasets. Finally, we highlight several current shortcomings and outline future directions.