Modern Subsampling Methods for Large-Scale Least Squares Regression

Subsampling methods aim to select a subsample as a surrogate for the observed sample. As a powerful technique for large-scale data analysis, various subsampling methods are developed for more effective coefficient estimation and model prediction. This review presents some cutting-edge subsampling methods based on the large-scale least squares estimation. Two major families of subsampling methods are introduced: the randomized subsampling approach and the optimal subsampling approach. The former aims to develop a more effective data-dependent sampling probability while the latter aims to select a deterministic subsample in accordance with certain optimality criteria. Real data examples are provided to compare these methods empirically, respecting both the estimation accuracy and the computing time.

Fuzzy Control in Cyber-Physical Systems

Controllers are devices regulating the operation of other devices or systems. Fuzzy controllers analyze the input data in terms of variables which take on continuous values in the interval [0, 1]. Since fuzzy logic has the advantage of expressing the solution of the problems in the natural language, the use of fuzzy instead of traditional controllers makes easier the mechanization of tasks that have been already successfully performed by humans. In the present paper a theoretical fuzzy control model is developed for the braking system of autonomous vehicles, which are included among the most characteristic examples of Cyber-Physical Systems. For this, a simple geometric approach is followed using triangular fuzzy numbers as the basic tools.

Credit Risk Assessment of Internet Financial Platforms Based on BP Neural Network

Aiming at the problem of credit risk, this paper selects key data indicators to establish an index system combining with the factors affecting the credit risk of the platform. Python crawler software was used to obtain relevant data of net lending platforms, and the crawled data of more than 1000 platforms were preprocessed. Ninety-five platforms with complete data were selected to build a BP neural network risk assessment model. The BP neural network model is used to make an empirical analysis of the risks of online lending platforms by using the data obtained, and the evaluation method of this paper is compared with the rating method of online lending sky eye. The empirical results show that the error of BP neural network can be stable at about 0.5, and the accuracy rate of evaluation is as high as 95.45%, which is much higher than the accuracy rate of 44.21% of net loan platform. This paper provides decision support for the credit risk early warning of net loan platform.

Atmospheric Boundary Layer Dynamics Evaluation Using Piezo-Resistive Technology for Unpowered Areal Vehicles

The paper presents a method for real-time observing of the convectional processes in the atmosphere boundary layer. The essence of the method is in providing real-time measurement of temperature, humidity, and pressure during the flight of a glider (soaring flight). Based on these measurements, a real-time evaluation of the atmosphere dynamics is presented. Measurements are taken during soaring flight of the glider and during the flight of a remotely controlled quadrocopter. Additionally, a method for atmosphere thermal identification by the measured parameters is introduced. The main application areas of this work are in unpowered flights, as well as in extending the flight time and distance of powered aerial vehicles. Moreover, the paper can be useful in research and observation of the lowest portion of the atmosphere and micro-scaled atmosphere dynamics evaluation.

A Low-Cost Experimental Testbed for Energy-Saving HVAC Control Based on Human Behavior Monitoring

Heating, ventilation, and cooling (HVAC) is the largest source of residential energy consumption. Occupancy sensors' data can be used for HVAC control since they indicate the number of people in the building. HVAC/sensor interactions show the essential features of a typical cyber-physical system (CPS). However, there are communication protocol incompatibility issues in the CPS interface between the sensors and the building HVAC server. Through either wired or wireless communication links, the server always needs to understand the communication schedule to receive occupant values from sensors. This paper proposes two hardware-based emulators to investigate the use of wired/wireless communication interfaces for occupancy sensor-based building CPS control. The interaction scheme between sensors and HVAC server will be discussed. The authors have built two hardware/software emulation platforms to investigate the sensor/HVAC integration strategies. The first emulator demonstrates the residential building's energy control by using sensors and Raspberry pi boards to emulate the functions/responses of a static thermostat. In this case, room HVAC temperature settings could be changed in real-time with a high resolution based on the collected sensor data. The second emulator is built to show the energy control in commercial building by transmitting the sensor data and control signals via BACnet in HVAC system. Both emulators discussed above are portable (i.e., all hardware units can be easily taken to a new place) and have extremely low cost. This research tests the whole system with YABE (Yet Another BACnet Explorer) and WebCTRL.

Exploring the Potential of an Extensible Domain-Specific Web Corpus for “Layfication”

This article presents experiments based on the extensible domain-specific web corpus for “layfication”. For these experiments, both the existing layfication corpus (in Swedish and in English) and a new addition in English (the NHS-PubMed subcorpus) are used. With this extended corpus, methods to classify lay-specialized medical sublanguages cross-linguistically using small data and noisy web documents are investigated. Sublanguage is a language variety used in specific domains. Here, the authors focus on two medical sublanguages, namely the “patientspeak” (lay) and the medical jargon (specialized). Cross-lingual sublanguage classification is still largely underexplored although it can be crucial in downstream applications for digital health and cyber-physical systems. Classification models are built using small and noisy training sets in Swedish and evaluated on English test sets. The performance of Naive Bayes classifiers—built with stopwords and with Bag-of-Words—is compared with convolutional neural network classifiers leveraging on MUSE multi-lingual word embeddings. Results are promising and nuanced. These results are proposed as a first baseline for cross-lingual sublanguage classification.

Architectural Modelling of Cyber Physical Systems Using UML

Cyber-physical systems (CPS) is an exciting emerging research area that has drawn the attention of many researchers. However, the difficulties of computing and physical paradigm introduce a lot of trials while developing CPS, such as incorporation of heterogeneous physical entities, system verification, security assurance, and so on. A common or unified architecture plays an important role in the process of CPS design. This article introduces the architectural modeling representation of CPS. The layers of models are integrated from high level to lower level to get the general Meta model. Architecture captures the essential attributes of a CPS. Despite the rapid growth in IoT and CPS a general principled modeling approach for the systematic development of these new engineering systems is still missing. System modeling is one of the important aspects of developing abstract models of a system wherein, each model represents a different view or perspective of that system. With Unified Modeling Language (UML), the graphical analogy of such complex systems can be successfully presented.

Fractional Order PIλDµ Control Design for a Class of Cyber-Physical Systems with Fractional Order Time-Delay models

The control of cyber-physical systems (CPS) is a great challenge for researchers in control theory and engineering mainly because of delays induced by merging computation, communication, and control of physical processes. Consequently, control solutions for time-delay systems can be applied efficiently for many CPS system configurations. In this article, a fractional order PIλ and PIλDµ control design is investigated for a class of fractional order time-delay systems. The proposed control design approach is simple and efficient. The controller parameter's adjustment is achieved in two steps: first, the relay approach is used to compute satisfactory classical PID coefficients, namely kp, Ti and Td. Then, the fractional orders λ and µ are optimized using performance criteria. Simulation results show the efficiency of the proposed design technique and its ability to enhance the PID control performance.

Data Dissemination for Vehicles in Temporary Cellular Network Dead Spots

The cellular network coverage in sparsely populated and mountainous areas is often patchy. That can be a significant impediment for services based on connections between vehicles and their environment. This article presents a method to reduce the waiting time occurring when a vehicle intends to send a message via a cellular network but is currently in a dead spot, i.e., an area without sufficient coverage. The authors introduce a data dissemination protocol that allows vehicles to connect through an ad-hoc network. The ad-hoc network peers can then find out which one will most likely leave the dead spot first. The selected vehicle stores then the messages of all connected vehicles and forwards them to the remote infrastructure as soon as it regains cellular network access. This research also discusses message flows in larger dead spots in which a vehicle may consecutively form several ad-hoc connections. Further, the authors describe an initial implementation of the protocol using the technology Wi-Fi Direct that is realized on most modern mobile phones.

An Image Processing and Machine Learning Approach for Early Detection of Diseased Leaves

India is largely an agriculture dependent country. It contributes to almost 17% of the GDP. A wide range of crops are grown throughout the year. Extensive cultivation also makes the plants prone to a lot of diseases. There are no efficient methods to detect these diseases from its outset. People in the rural areas where most of the agriculture happens are totally helpless in situations where most of their crops have been affected by disease. Most of the diseases that plague plants leave a characteristic feature on the leaf. By applying image processing techniques like image enhancement and feature extraction one can extract the required information required to analyze the type and severity of the disease. The obtained information when fed to a classifier like support vector machine (SVM), the plant can be classified to be affected by a certain disease. One can also determine the stage of the disease (infant or mid or terminal). Crop diseases impact the livelihood of those involved in agriculture immensely. Consumption of such produce also affects the health of humans and animals. Manually monitoring these diseases requires a lot of time and expertise. Hence, utilizing image processing for the detection of diseases is a better option. It takes into consideration the features which may not be determined visually. Consider the example of tomato crop in India which is prone to a number of diseases caused by pathogens, bacteria, viruses, and phytoplasmas-like organisms. Due to this disease the framers incur a huge loss. To overcome this problem a lot research is being conducted using image processing and neural network model for automatic detection of diseases using drone technology.