Autonomous Features Research Articles

Method of Voltage Sag Causes Based on Bidirectional LSTM and Attention Mechanism

The voltage sags’ cause recognition is the basis for formulating governance plans and clarifying liabilities for accidents. For voltage sag cause recognition methods based on physical characteristics, new challenges are presented in terms of accuracy, adaptability and algorithm efficiency. Deep learning is a method based on characterizing and learning data. The efficient mechanism of autonomous feature learning can effectively overcome the problems of information loss and generalization ability based on existing physical property methods. The long short-term memory network (LSTM) has the characteristics of memory and can better learn the data characteristics with time series characteristics. Bidirectional LSTM can consider historical information and future information compared with standard LSTM, and has more processing for time series data. While using the attention mechanism can highlight the key influencing factors in the time series and improve the recognition accuracy of the model. For the transient sag time series data, this paper proposes a multi-layer structure based on bidirectional LSTM and attention mechanism to classification recognition. The experiment uses simulation data and measured data to prove that the model has good recognition ability and good anti-noise performance in the recognition of voltage sag causes, and can be reliably applied in practical engineering.

Tesla, Inc.: Supercharging Auto Sales

This case describes the situation of Tesla, Inc. (Tesla), in mid-2019. Tesla had just successfully launched the more affordable Model 3, but sales of the most expensive and profitable Models X and S were stagnant. In order to incentivize the sales of these luxury models, Tesla recently brought back free Supercharging for life for customers purchasing Models X or S. The case provides information about the history of Tesla, one customer's hesitation about buying a Tesla electric vehicle (EV), purchasing prices, different ways to procure electricity and their cost, Supercharging stations, tax credits, and the autonomous and self-driving features of the car. Excerpt UVA-QA-0915 Jun. 10, 2020 Tesla, Inc.: Supercharging Auto Sales It was August 2019, and Henry Nugent woke up at 6:00 a.m. to go on a quick run before commuting to downtown Atlanta for his job at a regional investment bank. Reading the news on his smartphone in bed, Nugent froze when he saw the promotion he'd been waiting for: Tesla, Inc. (Tesla), was once again offering its popular free unlimited Supercharging for life on its Model S and Model X electric cars (Exhibit1). Nugent had debated buying the Model X for months. He was attracted to the roomy SUV, the falcon-wing doors, the car's immediate torque response, the panoramic glass, the regenerative breaking, and its smooth, silent ride. Nugent also liked the concept of reducing emissions. But the base price at $ 85,000, without including the $ 7,000 for the full self-driving capability, had made him think twice. He had also seen the reports that Models X and S sales were sluggish after the recent introduction of Model 3. In its early days, Tesla had offered free unlimited Supercharging as part of the package for buying one of its vehicles. By May 2017, the free Supercharging was used for a referral promotion where all existing Tesla owners could gift up to five friends free Supercharging. But in September 2018, Elon Musk, the cofounder and CEO of Tesla, ended the popular promotion with the explanation, “Sorry, it's not really sustainable at volume production & doesn't incent optimal behavior. We probably should have ended this earlier.” Instead, buyers of Model X and Model S were offered 1,000 free miles and then 5,000 free miles of charging. Tesla was also dangling cheaper insurance on its vehicles, along with the promise of a self-driving software upgrade. With the return of free unlimited charging, at least for a while, Nugent's interest peaked. How much was he really saving? Was free Supercharging the best way for Tesla to kick-start sales of Models X and S? . . .

Efficient Detection and Classification of Internet-of-Things Malware Based on Byte Sequences from Executable Files

Simple implementation and autonomous operation features make the Internet-of-Things (IoT) vulnerable to malware attacks. Static analysis of IoT malware executable files is a feasible approach to understanding the behavior of IoT malware for mitigation and prevention. However, current analytic approaches based on opcodes or call graphs typically do not work well with diversity in central processing unit (CPU) architectures and are often resource intensive. In this paper, we propose an efficient method for leveraging machine learning methods to detect and classify IoT malware programs. We show that reliable and efficient detection and classification can be achieved by exploring the essential discriminating information stored in the byte sequences at the entry points of executable programs. We demonstrate the performance of the proposed method using a large-scale dataset consisting of 111K benignware and 111K malware programs from seven CPU architectures. The proposed method achieves near optimal generalization performance for malware detection (99.96% accuracy) and for malware family classification (98.47% accuracy). Moreover, when CPU architecture information is considered in learning, the proposed method combined with support vector machine classifiers can yield even higher generalization performance using fewer bytes from the executable files. The findings in this paper are promising for implementing light-weight malware protection on IoT devices with limited resources.

A Miniature Robotic Turtle With Target Tracking and Wireless Charging Systems Based on IPMCs

State-of-the-art autonomous micro-robotic turtles suffer from various limitations, such as power restrictions that minimize their deployment times. In this paper, an Ionic Polymer Metal Composite (IPMC) actuator-based centimeter-level biomimetic underwater robot was designed and developed as a robotic turtle with self-charging capabilities to overcome such limitations. It could move forward and make turns driven by five IPMCs on the water. The underwater charging station was able to transmit wideband ultrasonic and electromagnetic fields for electromagnetic induction charging. An ultrasonic communication system with one ultrasonic transmitter and two ultrasonic receivers was first fabricated to implement communication between the underwater station and the biomimetic underwater robot for autonomous tracking and rechargeable capabilities. Experiments were carried out to confirm the operation of the biomimetic underwater robot, which verified the centimeter-level rechargeable capabilities and autonomous target tracking features. The micro-robot demonstrated a self-tracking radial displacement error of approximately 6 mm and a charging reliability rate of more than 73%.

ODBRANA POEZIJE: RUSKI FORMALIZAM I NOVA KRITIKA

The paper aims to present the key concepts that two formalist schools of thought developed in order to defend poetry as a miracle of communication. Russian Formalism takes Shklovsky’s defamiliarization (остранение) as its key concept and the trans-sense language (zaum) of Russian Futurist poets as a basis for analysing what constitutes differentia specifica of poetry. The ideas of the Russian formalist school, through Roman Jakobson, spread first to Eastern Europe, and then to the United States of America, where they influenced a group of critics, who were already inspired by T. S. Eliot’s and I. A. Richards’ ideas on poetic language and communication, to develop a new critical methodology. The name “New Criticism” was supposed to indicate that this school of thought was about different approaches and new tendencies in criticism. As the paper demonstrates, the key representatives of New Criticism are particularly interested in exploring the function of poetry and of criticism as well as the nature of poetic imagination and language. In order to examine what the poem says as a poem, they developed the practice of close reading and focused on metaphor, paradox, and specific method by which a poet transforms his experience into a poem as autonomous features of poetic expression. The special terminology introduced by Russian Formalism and New Criticism, and the complex, ironic and intellectual language they used not only managed to throw light on what specific problems of the science of literature were, but also enabled a defence through poetry – a kind of resuscitation and refinement of non-literary reality.

An Encrypted Speech Retrieval Method Based on Deep Perceptual Hashing and CNN-BiLSTM

Since convolutional neural network (CNN) can only extract local features, and long short-term memory (LSTM) neural network model has a large number of learning calculations, a long processing time and an obvious degree of information loss as the length of speech increases. Utilizing the characteristics of autonomous feature extraction in deep learning, CNN and bidirectional long short-term memory (BiLSTM) network are combined to present an encrypted speech retrieval method based on deep perceptual hashing and CNN-BiLSTM. Firstly, the proposed method extracts the Log-Mel Spectrogram/MFCC features of the original speech and enters the CNN and BiLSTM networks in turn for model training. Secondly, we use the trained fusion network model to learn the deep perceptual feature and generate deep perceptual hashing sequences. Finally, the normalized Hamming distance algorithm is used for matching retrieval. In order to protect the speech security in the cloud, a speech encryption algorithm based on a 4D hyperchaotic system is proposed. The experimental results show that the proposed method has good discrimination, robustness, recall and precision compared with the existing methods, and it has good retrieval efficiency and retrieval accuracy for longer speech. Meanwhile, the proposed speech encryption algorithm has a high key space to resist exhaustive attacks.

Coordinated UAV For Efficient Field Inspection

AbstractAirborne field inspection services conducted by traditional approach or by a single manual/autonomous Unmanned Aerial Vehicle (UAV), involves considerable operator work load in flight operations for coverage of very large inspection fields. There is a scope to reduce operational cost, time and operator work load for large scale inspection scenarios. This paper presents a system concept where the coordinated flight of many UAVs, as a Multi‐Agent‐System (MAS) may be used, such that, it is especially, “reconfigurable” based on the nature of inspection. This Multi‐Agent‐System concept is developed with autonomous features, such as self‐organization capabilities to adapt intelligently around the objects of inspection or during navigation. Also, the paper discusses on how coordinated flight contributes to improving the autonomy through iterative navigation planning and re‐distribution of work load into smaller sections, among the agents. Thus, the concept is expected to provide great benefit to the inspection service providers in terms of time, cost, and efforts while leveraging higher levels of autonomy.

Exploring Deep Physiological Models for Nociceptive Pain Recognition

Standard feature engineering involves manually designing measurable descriptors based on some expert knowledge in the domain of application, followed by the selection of the best performing set of designed features for the subsequent optimisation of an inference model. Several studies have shown that this whole manual process can be efficiently replaced by deep learning approaches which are characterised by the integration of feature engineering, feature selection and inference model optimisation into a single learning process. In the following work, deep learning architectures are designed for the assessment of measurable physiological channels in order to perform an accurate classification of different levels of artificially induced nociceptive pain. In contrast to previous works, which rely on carefully designed sets of hand-crafted features, the current work aims at building competitive pain intensity inference models through autonomous feature learning, based on deep neural networks. The assessment of the designed deep learning architectures is based on the BioVid Heat Pain Database (Part A) and experimental validation demonstrates that the proposed uni-modal architecture for the electrodermal activity (EDA) and the deep fusion approaches significantly outperform previous methods reported in the literature, with respective average performances of and for the binary classification experiment consisting of the discrimination between the baseline and the pain tolerance level ( vs. ) in a Leave-One-Subject-Out (LOSO) cross-validation evaluation setting. Moreover, the experimental results clearly show the relevance of the proposed approaches, which also offer more flexibility in the case of transfer learning due to the modular nature of deep neural networks.

Introduction to the Special Section: The Autonomisation of Weapons Systems: Challenges to International Relations

The collection of eight articles in this special section provides insightful thinking points in the context of the political debate, the academic conversation, and the public interest in novel security technologies with autonomous features. These articles are also a call for further, empirically and theoretically informed research into the implications of AI for the international security dimension, specifically, and for societies more generally.

Trustworthiness of unmanned automated subway services and its effects on passengers’ anxiety and fear

Commuters worldwide can now travel to work using unmanned subway services while tourists may travel to holiday destinations on planes that employ automatic flight control equipment. More recently, autonomous buses have made their debut in several cities. Technological advances in transport services, however, are accompanied by challenges to passengers’ perceived safety and the acceptability of the technology. While passengers’ anxiety and fear appear to reduce their acceptance of new technologies, trust has been shown to positively affect their acceptance of automation and autonomous vehicles. In this vein, the present study investigates the trustworthiness of automated subway services in Korea with regard to passenger anxiety and fear. First, automated subway-specific trustworthiness factors are identified through exploratory factor analysis. Subsequently, the effects of the identified factors on passenger anxiety and fear are examined using structural equation modeling. The results indicate that attributes that contribute to the trustworthiness of automated subways are grouped into automation- and service-related factors: ‘automation trustworthiness’ and ‘service trustworthiness’, both of which appear to negatively affect passengers’ anxiety and fear, mediated by the passengers’ sense of certainty and coping potential. The negative effects of trustworthiness on anxiety and fear indicate that strengthening of the trustworthiness factors must be accompanied with technological advances in transportation. The implications for strengthening the trustworthiness of unmanned automated subways and other transport services employing autonomous features are presented, correspondingly, in this study.

Abstract 1982: Sustained CXCR2 signaling promotes secretory phenotype in cancer-associated fibroblasts of pancreatic cancer via activation of NFkB

Abstract Pancreatic cancer (PC) is the fourth leading cause of cancer-related deaths in the US and remains one of the most challenging malignancies. Desmoplasia and tumor-supporting inflammation are hallmarks of PC. In addition to the autonomous aggressiveness feature of PC malignant cells, host tumor microenvironment contribute greatly to tumor progression and spread. Cancer-associated fibroblasts (CAFs), represent the major component of the tumor microenvironment in PC, are implicated in facilitating therapy resistance and metastasis. Recent reports emphasize the concurrence of multiple subtypes of CAFs that carry out different functions. CXCR2 is a chemokine receptor that is known for its role during inflammation by recruiting innate immune cells and the induction of angiogenesis. In PC, oncogenic Kras upregulates CXCR2 and its ligands (CXCL1-3 and 5-8), which are linked to inducing tumor proliferation and immunosuppression as well as induction of malignant cells stemness contributing to therapy resistance. Deletion of CXCR2 in a syngeneic mouse model (Cxcr2-/-) of PC increased the abundance of fibrosis revealing a potential undescribed role of CXCR2 in regulating CAFs. We hypothesize that CXCR2 regulates CAFs function in PC and contribute to CAFs heterogeneity. Co-culture of PC tumor cells in the conditioned media of CAFs and co-culturing CAFs in conditioned media of tumor cells revealed an oncogenic Kras-dependent interaction. CAFs enhanced the growth of PC cells with oncogenic Kras; whereas, conditioned media of the PC cells with oncogenic Kras has inhibited CAFs growth. Paracrine factors derived from tumor cells decreased the expression of the fibrotic CAFs-associated markers including αSMA and collagen I, and increased the expression of immunosuppressive cytokines and tumor-promoting chemokines including IL-4, IL-10, IL-13 and CXCL7. Utilizing recombinant CXCR2 chemokines and CXCR2 inhibitors, we confirmed the involvement of CXCR2 in this secretory phenotype. CXCL1 and CXCL8 exhibited a similar effect on CAFs by downregulating αSMA and upregulating IL-4, IL-13 and CXCL7. CXCR2 inhibitors reverted some of the observed effects. Examining downstream signaling pathways revealed that this effect is mediated through activation of NFκB. We further confirmed these findings in-vivo. Immunohistochemistry revealed higher αSMA expression in the tumors implanted in Cxcr2-/- compared to wildtype mice. Altogether, we demonstrate that sustained signaling through CXCR2 activates NFκB and induces a secretory phenotype of CAFs in PC that upregulates pro-tumor immunosuppression growth factors. Citation Format: Mohammad Awaji, Michelle Varney, Abhilasha Purohit, Lingyun Wu, Sugandha Saxena, Sushil Kumar, Surinder K. Batra, Rakesh K. Singh. Sustained CXCR2 signaling promotes secretory phenotype in cancer-associated fibroblasts of pancreatic cancer via activation of NFkB [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1982.

An Inquest into the Distinctive Features of Malaysian Military Law

Military forces all over the world have independent legal systems which are exclusively designed to warrant the smooth running of the armed forces administratively and to ensure utmost respect to the backbone of the military, namely discipline. Thus, when a person decides to be a member of armed forces, he should be mentally and physically prepared that he is going to be subjected to an additional law, namely military law. Anywhere in the world, this law is administered by a system which is independent of the standard civil legal system. This article is intended to discuss the unique characteristics of Malaysian Military Law in four (4) dimensions namely the threefold objectives of military law, the autonomous features of military law, the legality of command power, and the unique characters of Malaysian military judicial hierarchy system.

Lorentz Boost Networks: autonomous physics-inspired feature engineering

We present a two-stage neural network architecture that enables a fully autonomous and comprehensive characterization of collision events by exclusively exploiting the four-momenta of final-state particles. We refer to the first stage of the architecture as Lorentz Boost Network (LBN). The LBN allows the creation of particle combinations representing rest frames. The LBN also enables the formation of further composite particles, which are then transformed into said rest frames by Lorentz transformation. The properties of the composite, transformed particles are compiled in the form of characteristic variables that serve as input for a subsequent network. This second network has to be configured for a specific analysis task such as the separation of signal and background events. Using the example of the classification of tt̄tH and tt̄+bb̄ events, we compare the separation power of the LBN approach with that of domain-unspecific deep neural networks (DNN). We observe leading performance with the LBN, even though we provide the DNNs with extensive additional input information beyond the particle four-momenta. Furthermore, we demonstrate that the LBN forms physically meaningful particle combinations and autonomously generates suitable characteristic variables.

Algal cell response to laboratory-induced cadmium stress: a multimethod approach.

We examined the response of algal cells to laboratory-induced cadmium stress in terms of physiological activity, autonomous features (motility and fluorescence), adhesion dynamics, nanomechanical properties, and protein expression by employing a multimethod approach. We develop a methodology based on the generalized mathematical model to predict free cadmium concentrations in culture. We used algal cells of Dunaliella tertiolecta, which are widespread in marine and freshwater systems, as a model organism. Cell adaptation to cadmium stress is manifested through cell shape deterioration, slower motility, and an increase of physiological activity. No significant change in growth dynamics showed how cells adapt to stress by increasing active surface area against toxic cadmium in the culture. It was accompanied by an increase in green fluorescence (most likely associated with cadmium vesicular transport and/or beta-carotene production), while no change was observed in the red endogenous fluorescence (associated with chlorophyll). To maintain the same rate of chlorophyll emission, the cell adaptation response was manifested through increased expression of the identified chlorophyll-binding protein(s) that are important for photosynthesis. Since production of these proteins represents cell defence mechanisms, they may also signal the presence of toxic metal in seawater. Protein expression affects the cell surface properties and, therefore, the dynamics of the adhesion process. Cells behave stiffer under stress with cadmium, and thus, the initial attachment and deformation are slower. Physicochemical and structural characterizations of algal cell surfaces are of key importance to interpret, rationalize, and predict the behaviour and fate of the cell under stress in vivo.

Attention based residual network for medicinal fungi near infrared spectroscopy analysis.

As an effective technology, near infrared spectroscopy (NIRS) can be widely applied to analysis of active ingredients in medicinal fungi. Multiple regression methods are used to compute the relationship between spectral vectors and ingredient contents. In this paper, an autonomous feature extraction method by using attention based residual network (ABRN) to model original NIRS vectors is introduced. Attention module in ABRN is employed to enhance feature wave bands and to decay noise. Different from traditional NIRS analysis methods, ABRN does not require any preprocessing of artificial feature selections which rely on expert experience. The experiments test ABRN by analyzing original spectrums of medicinal fungi (Antrodia Camphorata and Matsutake), which are from 800 nm to 2500 nm, and predicting active ingredients within them. We compare ABRN with other popular NIRS analysis methods. The root mean square error of Antrodia Camphorata training set (RMSET) and validation set (RMSEV) are 0.0229 g⋅g⁻¹ and 0.0349 g⋅g⁻¹ for polysaccharide, and 0.0173 g⋅g⁻¹ and 0.0189 g⋅g⁻¹ for triterpene. The RMSET and RMSEV of Matsutake are 0.1343 g⋅g⁻¹ and 0.2472 g⋅g⁻¹ for polysaccharide, and 0.0328 g⋅g⁻¹ and 0.0445 g⋅g⁻¹ for ergosterol. The (coefficient of determination) of these four ingredients are 0.711, 0.753, 0.847 and 0.807. The results indicate that ABRN has better performance in autonomously extracting feature wave bands from original NIRS vectors, which can decrease the loss of tiny feature peaks.

The Autonomous Industrial Plant -Future of Process Engineering, Operations and Maintenance

Abstract Almost every day we read about new advances in self-driving cars and artificial intelligence. For autonomous driving, there already exist established standards with six levels describing the steps from driving without assistance to fully automated, i.e., autonomous driving. While many people and industries - in particular the automotive industry - believe that self-driving cars are feasible in the near future, one may ask how feasible full autonomy is in the process industry? Is it realistic to aim towards a fully autonomous chemical plant, and what does full autonomy even mean for the process industries? This paper introduces a definition of autonomy and corresponding levels of autonomy for the process industry. We start from an abstract definition of six autonomy levels that are applicable to various industrial domains, also illustrated by recent examples from different industrial applications such as mining or marine. Then, we discuss the applicability of this definition for process industry in a first step. In a second step of refinement, we also study some key autonomous features of process operations and give examples for the meaning of each autonomy level per feature.

Autonomous feature type selection based on environment using expectation maximization in self-localization

Self-localization in autonomous robots is one of the fundamental issues in the development of intelligent robots, and processing of raw sensory information into useful features is an integral part of this problem. In a typical scenario, there are several choices for the feature extraction algorithm, and each has its weaknesses and strengths depending on the characteristics of the environment. In this work, we introduce a localization algorithm that is capable of capturing the quality of a feature type based on the local environment and makes soft selection of feature types throughout different regions. A batch expectation–maximization algorithm is developed for both discrete and Monte Carlo localization models, exploiting the probabilistic pose estimations of the robot without requiring ground truth poses and also considering different observation types as blackbox algorithms. We tested our method in simulations, data collected from an indoor environment with a custom robot platform and a public data set. The results are compared with the individual feature types as well as naive fusion strategy.

A miniature integrated nuclear reactor design with gravity independent autonomous circulation

Abstract A miniature integrated nuclear reactor design with gravity independent autonomous circulation (ACMIR) was newly proposed. The reactor core, energy transfer system of Stirling and linear electric motors are integrated in the reactor pressure vessel to achieve high power density and autonomous circulation capability. The coolant circulation is autonomously driven by gas expansion at heat end and compression at cool end thus is independent on gravity. Twelve sets of rotary drum controller and reflector are used to regulate reactivity outside the reactor pressure vessel. The physics and thermodynamic properties, as well as the safety performances are analyzed. According to these analysis, inherent safety characteristics are obtained, and the reactor is capable to shut down and remove residual heat passively without any external intervention, in accident conditions such as loss of external power supply, overtemperature/overpressure of the reactor, impact when rocket launching or landing, stagnation of the displacer or piston, et al. The integration, less pipes design, gravity independent autonomous circulation features make it a good candidate for space flight propulsion, the Moon and Mars base power supply, the deep-sea or other tilt and swing situation applications.

Concepts of Teleoperated Robots

This paper provides basic overview of teleoperated robotic systems and their characteristics. The aim of the paper is not to describe such robotic systems in detail such as suggesting actual hardware and software, but to provide basic information, which has to be considered during the design and usage of such robot. Paper is focused on general description of functionality, characteristics and applications of teleoperated robots. Overview of potential use cases is provided with description of potential downfalls. Means of communication with the robot are discussed and elaborated with respect to reliability and safety. Available types of human to machine interfaces and sensor systems are listed and described considering reliability, safety and quality of control. Especially safety issues are highlighted, because such robots suffer from the involvement of humans in the control scheme. This paper comes to a conclusion that it is of primary importance to consider safety measures and stresses vulnerability of whole system introduced by operator and means of communication. In this paper is further stressed that in case of lost communication teleoperated robot must be able to either safely continue its function or stop. Future research will be focused on improving teleoperation via virtual reality, further improvement of autonomous features leaving operator only to decide which actions to perform and cloud computing.

A Platform for Developing Robotic Navigation Strategies in a Deformable, Dynamic Environment

While in vivo robotic locomotion has demonstrated its potential through a number of means, sensing and closed-loop control for such devices have not been at the forefront of research. By enabling real-time navigation, capsule endoscopy could transform from passive capsule endoscopes (CEs) and remote-controlled robotic CE devices into semi- or fully autonomous, surgical tools. In this work, a novel robotic endoscope platform (REP) is designed and tested. This platform will be used to evaluate control, localization, and mapping techniques necessary to endow a robotic CE with autonomous features. REP performance was benchmarked against design requirements that included sensing, speed, and maneuverability. It was found that the REP could accelerate to a maximum speed of 80 mm/s in less than 0.2 s and could achieve a zero-point turning radius. Most significantly, a low-cost image segmentation algorithm was developed to enable real-time lumen navigation/center tracking. Visual navigation time trials were compared to manual user trials through a 90-degree turn in a deformable simulator, exhibiting mean run times 34% faster than manual users. This demonstrates the REP's maneuverability and sensing capabilities allowing for future implementation of a range of advanced sensing and control techniques ahead of robotic CE adoption.