Dorina Purcaru*, Anca Purcaru**
* Department of Automation and Electronics, University of Craiova Craiova, Romania (e-mail: dpurcaru@electronics.ucv.ro). **VIG IMPEX Craiova, Romania (e-mail: aipurcaru@gmail.com)

Abstract: An efficient training in the measurement systems field requires theoretical knowledge and practical skills in using sensors, transducers, electronic devices and circuits. The paper presents a guideline for thorough understanding the operation and programming of inductive, capacitive and ultrasonic sensors, resistance temperature detector and intelligent temperature sensor. The information from user manuals and datasheets is selected, explained and completed with suggestive graphic representations and static transfer characteristics. Three experimental systems allow the sensors' testing and monitoring and implement a temperature control system and a motion control system. The sensor monitoring and remote programming and the applications' implementation require a programmable logic controller. These experimental systems have been proved very useful in academic study for experimental training on measurement systems.
Keywords: education, sensors, monitoring, testing, programmable logic controller, programs.

Laviniu Aurelian Badulescu
Faculty of Automation, Computers and Electronics, University of Craiova, Craiova, Romania (e-mail: badulescu@automation.ucv.ro)

Abstract: Decision Trees are some of the most well-liked Data Mining algorithms for prediction and classification. We will present in this paper the performance of the Decision Trees (DT) induced with 13 splitting criteria over Census Income database taken from UCI Knowledge Discovery in Database Archive. The research involved the growing of the DT on a training data set, the extraction of decision rules, the pruning of the DT using two pruning methods: error-based and pessimistic and lastly, the DT execution on the test data. A good selection of an attribute used to split the data set at each DT node will optimize the classification accuracy. The biggest values for the classification accuracy are systematically obtained by error-based pruned DT. In the same time, the best performances for classification error rate on the test data set, data absolutely not known at the training of DT, were obtained by the symmetric gini index splitting criterion. We conclude the paper with a discussion on related work.
Keywords: decision trees; classification error rate; splitting criterion; pruning method; decision rule.

Faiq Baji*, Mihai Mocanu**
Computer Science and Information Technology Department, Automation, Computers and Electronics Faculty, Craiova University, Craiova, Romania, *e-mail: faiq.baji2015@gmail.com **e-mail: mmocanu@software.ucv.ro

Abstract: In recent years, there has been an obvious increase in the number of melanoma skin cancer patients. The diagnosis of skin cancer at an early stage is very important for effective treatment, but it is a challenging task in melanoma diagnosis. The irregularity of the lesion border is an important tools, it is used to reporting the melanoma during growth and propagation. In this paper, a new method is proposed to calculate Boundary Irregularity; this method is based on dividing the lesion according to the major axis into two sections, within each section, the Euclidean distance between the centre of the lesion and all pixel's borders is computed. The proposed method is based on the ABCD rule of dermatoscopy, statistical texture analysis, and lesion symptoms for melanoma skin cancer diagnosis. The computation of ABCD rule and texture analysis methods are based on the connected components objects for image segmentation. Several approaches have been introduced to identify the geometric shape features such as compactness, dispersion, and border irregularities. In this paper, the Stolz algorithm has been used for the diagnosis by calculating the Total Dermatoscopy Value (TDV) score. As it can be seen from the experimental results, the Boundary Irregularity method with the proposed method shows acceptable performance rates, the accuracy of the system is 90%
Keywords: Melanoma, Skin Cancer, Connected Components Objects, ABCD Rule, Texture Feature

Traian-Titi ?erban
University of Craiova (e-mail: tserban@electronics.ucv.ro)

Abstract: This paper presents a computerized instrument for temperature data logging using multiple sensor types. A versatile SCADA-like system able to display time sequences of the temperature evolutions of many measurement points is described. Based on a modular structure, the system allows for the temperature measurement using any of the main temperature sensors: thermocouples, RTD and analog output PTAT sensors. A specialized data acquisition microcontroller is used to convert the analog signals into digital data to be transmitted to a host computer that exhibits the user interface and displays the evolution of the temperature. The described system is dedicated to environment monitoring systems and to any industrial processes involving slow evolutions of temperature.
Keywords: temperature measurement, sensors, data logging, wireless data transmission

George-Cristian Calugaru*, Ionut Cristian Resceanu*, Elena Andreea Danisor*
*University of Craiova, Faculty of Automation, Computers and Electronics (e-mails: calugaru.george.nds@gmail.com, resceanu@robotics.ucv.ro, andreea_elenadanisor@yahoo.com)

Abstract: This article proposes an extension of the feasible sets theory in aircraft constrained optimal control problems. After a brief introduction in the constrained optimal control problem, its objective and the feasible solutions analysis, a case study for the proposed extension is presented which employs a model of an aircraft during the last phase of landing. Taken into account the specific parameters of model in the case study and for a given prediction horizon, the explicit controllers determined at each step together with the regions of the feasible partitions. Analyzing the feasible partitions under different use cases can present interest in design and development of aircraft systems which also employ model predictive control.
Keywords: aircraft control, constrained optimal control, model predictive control

Viorel Costin Ban?a, Dorian Cojocaru
Mechatronics and Robotics Department, Faculty of Automation, Computers and Electronics, University of Craiova, Romania (e-mail: bantaviorel@gmail.com)

Abstract: The paper deals with what the Internet offers in order to connect our ERP systems, using mobility and SAP Fiori interface, with actual business context. With this article we want to covered some aspects in order to show how metrology principles are integrated with an Enterprise Resource Planning (ERP) systems, how we can have utilized the interface SAP Fiori - for mobility purposes, through enabled learning organization and software as a service. The authors have been focused to describe, discuss and show the methodology of development of a metrology software application included in e-metrology category for the calibration of oil tanks. For this have been used the new technologies from SAP - WebDynpro, SAP Fiori interface, implemented and tested in a gas and oil business environment. The existing ERP - SAP in fact, which is the environment for the application, is installed in the oil and gas company, this means that our application is integrated with all processes existed in area of Demand Management environment. The development / programming of the CRT application was done using the ABAP programming environment, using WebDynpro technology as a Web-oriented programming method. All necessary phases (analysis, testing and implementation) were part of the entire development agreed methodology for our software CRT tool. The CRT application for calibration management was tested in the E&P and Marketing divisions of an oil and gas company, this article highlighting the extension to the mobility of the application using the SAP Fiori interface.
Keywords: SAP Fiori interface, SOA, SaaS, e-Metrology System, ERP III, Enabled Learning Organizations, Calibration, ERP.