Procesare distribuita
Editura Universitatii Suceava, 2007
ISSN 1842-6808
Distributed Processing
Table of Contents
MODELLING GESTURES AND GESTURE DICTIONARIES FOR HUMAN COMPUTER INTERACTION
Radu-Daniel VATAVU1, Stefan-Gheorghe PENTIUC2
Universitatea "Stefan cel Mare" Suceava
Str. Universitatii 13, 720229, Suceava, România
1vatavu@eed.usv.ro,2pentiuc@eed.usv.ro
Abstract.
The use of gesture as means of interacting has thus been very rapidly proliferating and many systems have emerged that allow users to touch, point, work and travel by means of head, hand, arm, gaze or whole body gestures. The most common areas of application are in sign language recognition, gesture to speech, virtual environments, 3D modeling, human robot dialogue, and the list goes on.
In this paper we follow human gesture in all its representations as it appears at different levels in the HCI domain (such as raw data streams for the low-level tasks of gesture acquisition, mathematical model for analysis, pattern for recognition, record for database / dictionary of gestures, event in a high end programming language.) We provide a mathematical formalization for gestures in order to develop further discussions on gesture dictionaries. With regards to gesture as command, we identify and provide models for four different types of gestures by taking into account both posture information (or static dimension of gesture) and underlying motion trajectories (dynamic component).
Interaction is performed using both motion trajectories (dynamics) and postures (or static information). We may thus provide the following classification of gesture commands with regards to their structural pattern (amount of posture / motion considered for performing the command): static simple gestures, static generalized gestures, dynamic simple gestures and dynamic complex gestures.
The complex nature of gesture recognition makes human gestures have different representations at different levels. We have identified four distinct processing layers corresponding to the actual gesture execution in the real world, acquisition process, modeling and application. We follow the gesture in all its representations starting from the raw format as provided by an acquisition device to the final actual interaction stage (where gesture triggers action) at the highest application level.
Keywords:
gestures, gesture representation, dictionary of gestures
INFORMATICAL SYSTEMS FOR ASSISTING THE SPEECH THERAPIST IN ACCOMPLISHING THE PERSONAL THERAPY FOR SPEECH DISORDERS
Iolanda TOBOLCEA
Universitatea "Alexandru Ioan Cuza" Iasi
Str. Toma Cozma nr. 3, Iasi, Romania
itobolcea@yahoo.com
Abstract.
The progress achieved in computers science and engineering contribute to designing and implementing of rapid efficiency systems into the speech therapy, so the computer became a necessary and real clinical instrument also in the domain of speech disorders. The speech therapy specialist can use in harmony his clinical and technical abilities in order to help his patient, in this way the direction is more flexible and modern for evaluating, treating and for the constant feedback.
The usage of audio-visual equipments in the Romanian speech therapy was made in some cases of speech disorders, such as: Echophone – electronic device which delays the audio feedback (DAF), equipment proposed and experimented by I. Tobolcea (1995), Software LOGOPED 1.0 – software designed and experimented by I. Tobolcea (2000) for stuttering treatment., Software DISLEX-TEST - software designed under the coordination of I.Tobolcea (2005) which has the main objective gradating the similar optic letters: m-n; u-n; p-b; b-d; isolated, in words and in sentences.
Project „TERAPERS – Informatical system for personal therapy of speech disorders” was proposed in the idea of developing softwares which can help the Romanian speech therapists. This project started in July 2006 and it will be finalized in August 2008, partnership between the University „Stefan cel Mare” Suceava, University „Al.I.Cuza” Iasi and the Technical University „Gh.Asachi” Iasi.
The main objectives of the project team from the University “Al.I.Cuza” Iasi is the research course in the purpose of establishing the evaluation methodology and diagnostic of speech disorders, to elaborate the therapeutic guide and the exercises set for development of an expert system for personal therapy of speech disorders, which can allow the accomplishment of a training path for individual articulation depending on defect, antecedent experience in articulation therapy and antecedent evolution of the child.
The final product of the project will be represented by a smart system installed on the computer from the speech therapist office, including also an audio-visual model of the articulation mechanism and a system designed to be a therapy “friend” of the child.
At the international level there are a large number of therapeutic softwares, such as: Project Ortho-Logo-Paedia, Laureate Learning Systems, Speech Training, Assessment, and Remediation System, etc.
We can estimate that this research will stimulate the preoccupation for the classical therapy efficiency compared with the therapy assisted by computer, at national and also European level. The after results will be compared with the similar smart systems from other countries which use a similar phonetic language to the Romanian language.
Keywords:
therapy, computer, speech disorders, child, software, audio-visual
DISTRIBUTED INFORMATION SYSTEM FOR THERAPY OF DYSLALIA
Mirela DANUBIANU1, Stefan-Gheorghe PENTIUC2, Ovidiu SCHIPOR3, Cristian BELCIUG4, Ioan UNGUREANU5, Marian NESTOR6
1,2,3,4,6Universitatea "Stefan cel Mare" Suceava
5GENPRO Iasi
1mdanub@eed.usv.ro
2pentiuc@eed.usv.ro
3schipor@eed.usv.ro
Abstract.
The informatic system for dislalia therapy is part of the project „System for personalized therapy of speech disorders – TERAPERS”, financed through the 56-CEEX II03 /27.07.2006 contract, system that is meant to assist the specialists in speech therapy.
The main purpose of the system is : the initial and during therapy evaluation of the sampled children and the identification of some standard methods of their progresses or retrogression (taking into account the level of physiological and behavior parameters), the formalization of the evaluation methodology of the children, and the deployment of a knowledge data base in this area, the development of an expert system for the speech therapy that could allow an personalized therapy program for pronunciation training, depending on the speech disorder, the design and development of a mobile device intended for children use in assisted therapy which can offer feedback according to their age, to the solved exercises and to the results obtained , and the integration of the system for the optimal interaction between the child and the speech therapist.
The system is made of 2 parts: one installed on the computer in the speech therapist’s office and another one on the mobile device.
The fix component can manage tasks as: the therapists assistance in children complex evaluation for trouble discovery or the organization of the therapy the results analysis. The mobile device consists of sets of personalized exercises and can give a feedback of individual activities.
Keywords:
speech therapy, mobile device, expert system
SYSTEM MONITOR FOR SPEECH THERAPY ASSISTENCE
Mirela DANUBIANU1, Marian NESTOR2
Universitatea "Stefan cel Mare" Suceava
1mdanub@eed.usv.ro
2mnestor@stud.usv.ro
Abstract.
Speech is the most used means for the interhuman communication. It has an individual character manifested at the physiological and at the psychological level.
The speech disorders are manifested by the difficulties of some sounds pronunciation (dislalia), by the replacement of some sounds with another sounds, difficulties of writing and stuttering. An often-met type of troubles refers to pronouncing and articulation (dislalia, disartria, etc). They have a high rate especially with children, and dislalia is more often seen at very young persons. The symptoms for dislalia are sounds, syllabus and often-even word deformation, omission, substitute, replacing or reversing.
The speech therapist intervention means to identify specific objectives like: to find, completely evaluate and identify the speaking and communication trouble at pre-scholar children within kindergarten or not and at little scholars, the insurance of preventive character of the speech therapy by according a special attention to pre-scholars having speaking difficulty with the purpose of preventing the installation of learning difficulty of reading and writing when school should start, the orientation of speech therapy towards correction, recover, compensation, adaptation and school and social integration, and the correction of the speaking and communication trouble in order to avoid the risk of scholar failure
The information society leads to a new stage of human civilization development, to a superior life style. In this context it is impossible to ignore the impact of information technology concerning on improvement of methods for assistance of the persons with different kinds of disorders.
LOGOMON is a part of the intelligent system for the individually therapy of the speech disorders - TERAPERS. It is a software module that performs the specific tasks of a monitor system. These tasks are: the initial evaluation for hunting out the children with speech disorders, their registration in the database, the propose of a possible diagnostic for these children with the possibility for the expert to confirm or to modify this diagnostic, the selection of children for the therapy, the management of the therapy process and the supervision of the children evolution.
Moreover by using LOGOMON we collect and process the data such as we can remove any manual action of collecting or processing of data. Also we can eliminate the data storage on the paper.
The combination of the database technology with the advanced programming languages and with diverse signals recording and signals processing techniques has allowed the designing and the achievement of this kind of system for monitoring of children with speech disorders.
Keywords:
speech disorders, intelligent system, diagnose, survey
AUTOMAT PARSING OF AUDIO RECORDINGS. TESTING CHILDREN WITH DYSLALIA. -THEORETICAL BACKGROUND-
Ovidiu-Andrei SCHIPOR1, Marian NESTOR2
Universitatea "Stefan cel Mare" Suceava
1schipor@eed.usv.ro
2mnestor@stud.usv.ro
Abstract.
In this paper we present our researches regarding automat parsing of audio recordings. These recordings are obtained from children with dyslalia and are necessary for an accurate identification of speech problems. We develop the ADM algorithm and we analyze the complexity of this solution. The main objectives of this task are:
- recording 120 children (60 with correct pronunciation and 60 with dyslalia)
- we must permit different audio environments during recording (some phonemes will be used for training a real time recognition system)
- the cost of recording devices and the children’s impact must be minimized
- after recording is necessary to split the stream into phonemes
- the speech therapist’s voice must be ignored
We utilize a digital voice recorder in High Quality mode and with VCVA (Variable Control Voice Actuator) activated. The record format is IMA-ADPCM, 16KHz and 4bits (16 bits PCM). A microphone was placed at 10 cm from mouth in order to minimize environment noise.
A software set of classes (C#) was created for handling audio stream (read, conversion between different format, write). We also propose an original solution for placing markers in audio stream. These markers are needed for a correct parsing af full recoding.
Keywords:
automat parsing, ADPCM, speech therapy, dyslalia, phoneme recording
ARCHITECTURE OF A FUZZY EXPERT SYSTEM USED FOR DETERMINATION THE THERAPY OF DYSLALIC CHILDREN
Stefan-Gheorghe PENTIUC1, Ovidiu-Andrei SCHIPOR2, Maria-Doina SCHIPOR3
Universitatea "Stefan cel Mare" Suceava
1pentiuc@eed.usv.ro
2schipor@eed.usv.ro
3vmdoina@yahoo.com
Abstract.
In this paper we present an architecture of a fuzzy expert system used for therapy of dyslalic children. With fuzzy approach we can create a better model for speech therapist decisions. A software interface was developed for validation of the system.
The main objectives of this task are:
• personalized therapy (the therapy must be in according with child’s problems level, context and possibilities)
• speech therapist assistant (the expert system offer some suggestion regarding what exercises are better for a specific moment and from a specific child)
• (self)teaching (when system’s conclusion is different that speech therapist’s conclusion the last one must have the knowledge base change possibility)
Fuzzy logic has ability to create accurate models of reality. It’s not an “imprecise logic”. It’s a logic that can manipulate imprecise aspects of reality. In the latest years, many fuzzy expert systems were developed.
For fuzzy rules and variables representation we utilize FCL (Fuzzy Control Language) standard.
In order to validate expert system, was developed a software interface. We also create rules for 8 input variables, 4 intermediate variables and 3 output variables.
Further work is related from knowledge base extension and expert system integration into LOGOMON application.
Keywords:
fuzzy expert system, speech therapy, FCL language, fuzzy variables, fuzzy logic
EXERCISES FOR CHILDREN WITH DYSLALIA.-SOFTWARE INFRASTRUCTURE-
Cristian-Eduard BELCIUG1, Ovidiu-Andrei SCHIPOR2, Mirela DANUBIANU3
Universitatea "Stefan cel Mare" Suceava
1belciugc@yahoo.com
2schipor@eed.usv.ro
3mdanub@eed.usv.ro
Abstract.
In order to help children with dyslalia we create a set of software exercises. This set has a unitary software block (data base, programming language, programming philosophy). In this paper we present this software infrastructure with its advantage and disadvantage.
The exercises are part of a software system named LOGOMON. Therefore, besides horizontal compatibilities (between exercises) appear vertical compatibilities (with LOGOMON system).
Concerning database tables used for modulus of exercises, a part of them is “inherited” from LOGOMON application and another are specific for exercises application. We also need to specify that there were necessary minor changes of database tables used by LOGOMON.
As programming language we used C#, implemented in Visual Studio 2005. We developed specific interfaces elements and classes. We also used multimedia resources that were necessary for exercises (images, correct pronouncing obtained from speech therapist recording, video clips). Another section of this application is related to loading of exercises on mobile devices (Pocket PC).
A part of code has been imported directly, but there were a lot of files that need to be rewritten. Anyway, the multimedia resources were used without any processing.
Keywords:
exercises for dyslalia, software architecture, multimedia database, speech therapy
PHONEMATIC EXERCISES FOR CHILDREN WITH DYSLALIA.-SOFTWARE APPLICATION-
Stefan-Gheorghe PENTIUC1, Cristian-Eduard BELCIUG2, Doina Maria SCHIPOR3
Universitatea "Stefan cel Mare" Suceava
1pentiuc@eed.usv.ro
2belciugc@yahoo.com
3vmdoina@yahoo.com
Abstract.
In order to help children with dyslalia we created a set of software exercises. In this paper we present only phonematic exercises together with a consistent user guide.
In development of exercises we based on “Therapeutic guide” that has been made available by researchers team lead by Associate Professor PhD. Iolanda TOBOLCEA from „Alexandru Ioan Cuza” University, Iasi.
There has been followed some psychological and speech therapist principles:
- the specific of children age
- set of rules concerning exercises
In this article we refer only to phonematic hearing exercises (in these exercises we tested the capacity of children to distinguish and differentiate phonemes):
- to identify words that contain certain phoneme
- to identify the word that does not contain certain phoneme from a set of words
- to distinguish certain phoneme from pairs of paronyms (a word from pair contain certain phoneme and the other word contain a similar phoneme)
There has appeared some difficulties because not all the words have an equivalent in image format (verbs, adjectives, adverbs).
Therefore correct pronouncing of words has obtained by recording in a studio for a good understanding.
Keywords:
exercises for dyslalia, phonematic exercises, speech therapy, speech recording
MULTIAGENT SYSTEM ARCHITECTURE FOR INFORMATION RETRIEVAL IN B2B SYSTEMS
Felicia GIZA1, Cristina TURCU2
Universitatea "Stefan cel Mare" Suceava
1felicia@eed.usv.ro
2cristina@eed.usv.ro
Abstract.
This paper contains an architecture of an information retrieval system that use the advantages offered by mobile agents to collect information from different sources and bring the result to the calling user. Multiagent system architecture has both mobile and static agents. Static agents manage the resources and facilitate access to them for mobile agents. Mobile agents migrate between the servers of the system and use the resources for accomplishing his tasks. Mobile agents are main components of the architecture that can migrate among the platforms of the associated companies and help the users to gain access to the desired information. The paper describes also the designing of a system that will be used for determine the traceability of a product and also for searching information about a specific entity. In this system, the mobile agent will have a general interface for specifying the parameters of search. For the searching mode, it can be specified the servers that should be visited. From the platform point of view, we decide that JADE platform is the best suited for our application because of his particularities (respects the agents standards, has support for both Linux and Windows operating systems and also has an extension for mobile devices (PDA, mobile phones). There can be possible two forms of searches: general search, where the itinerary of the agent is established based on optimal way algorithms, and particular search were the itinerary is previously established and it can not be changed. The second one is used for determining the traceability of a product. Beginning with the destination of a product, the mobile agent visits all the companies that had the product some time in their stock, in a singular way.
Keywords:
multiagent system, mobile agents, B2B, information retrieval
A LOW COST RFID BASED SOLUTION FOR TRACKING AND TRACING PRODUCTS ALONG THE SUPPLY CHAIN
Valentin VLAD1, Cristina TURCU2, Alexandru GOLOCA3
Universitatea "Stefan cel Mare" Suceava
1vladv@eed.usv.ro
2cristina@eed.usv.ro
3alexg@eed.usv.ro
Abstract.
RFID (Radio Frequency Identification), a technology existing for years, has potential uses in a variety of applications. Though not without issues and challenges, RFID is a promising technology which analysts expect to become ubiquitous in the coming years, helping organizations solve problems in supply chain management, security, personal identification, and asset tracking.
Track and trace identification systems allow companies to keep the trace of products from raw materials to sellable products, through distribution, and right to the end user. Although the concept is not a new one in many industries, recent initiatives have put new pressures on suppliers to uniquely identify their products starting in the manufacturing plant and out to the end user — with the goal of gaining real-time visibility to the process, product and resources.
The purpose of this paper is to present a low cost RFID based solution in tracking and tracing products along the supply chain.
The solution is represented by a control gate realized as a hardware/software ensemble, developed around a low cost microcontroller from ARM7 family and using as operating system the real-time kernel ARTX from Keil. The control gate may:
• Communicate with a RFID reader for reading and writing data from or to a transponder;
• Save the data read from a transponder into a flash memory;
• Generate events according to data read from transponders;
• Send generated events to a server (e.g. an OPC server), when it requests them;
• Read/Set digital I/O.
These devices may be used successfully in a distributed system for supply chain management, but also, with some software changes may be used in other RFID based applications.
Some reliable qualities of this type of control gates may be mentioned:
• Autonomy: the system is equipped with an accumulator assuring its functionality for a period of time, in absence of the power supply.
• Real time functionality: the operating system ARTX is a robust real-time kernel
• Extendibility: the Modbus protocol, used for communication, allows extension of the system by interfacing LCCG with new types of devices using the same protocol.
• Reduced size
This paper contain an overview of the RFID technology and its applicability in the supply chain, then is presented the hardware and software architecture of the control gate developed. Finally is presented other similar solution from Feig.
Keywords:
RFID, track and trace products, supply chain, Modbus, embedded devices, OPC.
TECHNOLOGICAL PLATFORM FOR THE INFRASTRUCTURE OF INTELLIGENT STORE WITH MOBILE ACCESS FACILITIES - INTELSHOP
Carmen Eleonora STAN1, Ioana Catalina POP2, Andrei IANCU3, Eugen POP 4, Emanuel ANGHEL5, Doru E. TILIUTE6
1SC ITC INSTITUTE FOR COMPUTERS SA / Microsystems and software applications
2The Company for Research, Development, Engineering and Manufacturing for Automation Equipment and Systems - SC IPA SA
3The National Institute for Research and Development in Informatics
4Universitatea "Stefan cel Mare" Suceava
1carmen.stan@itc.ro
4dtiliute@seap.usv.ro
Abstract.
Economic, demographic and technological changes impose the need to reconsider the traditional way of trading.
This paper presents the solution adopted within a project that aims to create an innovator technological infrastructure for an intelligent store, integrating various technologies: RFID (for the automated product identification) or RFID-enabled payment interface, GPS, GIS (for locating customers or the store and starting proximity alerts), wireless network communication, mobile access at information (on PDA or SmartPhone), data analysis integrated media (Business Intelligence) regarding customers and stocks management.
This technological platform will be the support of new customers and store personnel directed services.
Keywords:
RFID, GPRS, GIS, intelligent shop, mobile access
GEOGRAPHICAL INFORMATIC SYSTEM FOR DEGRADATED INTRAVILAN AREAS MONITORING OWING-TO THE INCONTROLLABLE WASTE STORING
Nicolae MORARIU1, Paul PASCU2, Sorin VLAD3, Valeriu LUPU4
Universitatea "Stefan cel Mare" Suceava
1nmorariu@eed.usv.ro
2paulp@seap.usv.ro
3sorinv@seap.usv.ro
4valeriul@seap.usv.ro
Abstract.
The main objective of the system is the monitoring of the affected areas by the incontrollable waste storing using the newest informational and communication technologies through the elaboration of a geographical positioning electronic system GPS/GIS.
Through automatic collecting of the environmental, meteorological and ecological data, there will be realized a monitoring system, equipped with sensors and/or traductors, which are able to measure and translate (into electrical signals) measures with meteorological character (such as: solar radiation intensity, temperature, humidity etc.), but also state measures of the ecological system (examples: the concentration of water and soil nutrients, air, soil and water pollutants, biomasses). The measures obtained will have to be converted into numerical values that will be stored into the system database. Taking into account the periodical manner of the environment data acquisition (at different time intervals), the system database is a temporary one, which means that, along the current data, the database contains also historical data, and the time factor (attribute) has an essential role.
Inside the telematic system for the on-line management of the degraded intravillan areas are defined the following data categories:
- Data concerning the waste management (monitored intravillan areas, waste, pollution sources, waste storehouses, waste processing stations);
- Data concerning the environment protection (environment quality parameters: water, soil, air);
- Spatial data (thematic maps).
Related to the processing mode, the spatial data are grouped into: descriptive data (textual) and graphical data. The descriptive data are organized into tables and the graphical data defines vector images through points defined by coordinates and outline descriptions. The spatial data are represented and can be shown on layers.
Another data category is constituted of raster images representing image files obtained through cadastral and topographical maps scanning, which can constitute graphic data sources obtained by vectorizing operation.
The system database is a distributed on the following levels:
- the central database server (SQL server Microsoft);
- the local database servers (Microsoft ACCESS).
The organization, description and processing of the spatial data request the use of a GIS type product. There are available many market solutions such as: ArcINFO from ESRI, AutoCAD Map from Autodesk and GEOMEDIA from INTERGRAPH which are products made in U.S.A. at a relative low price. There are also solutions made in Romania (by example: NetSET from Data Invest and MapSys from Geotop) which have a more accessible price.
Keywords:
monitoring, waste storage, degraded intravillan areas, thematic maps, GPS/GIS
THE NEW ORIGIN OF SWEEPINGS ADMINISTRATION
Valeriu LUPU1, Catalin LUPU2, Nicolae MORARIU3
Universitatea "Stefan cel Mare" Suceava
1valeriul@seap.usv.ro
2catalinlupu@seap.usv.ro
3nmorariu@eed.usv.ro
Abstract.
In this article i will present a new concept taking in consideration the sweepings administration and which consist in:
• the employment of neuronal network for predict serial time representing sweepings amount which are collected from population in time. The problems of prediction the serial time can be associated with neuronal network call “feed-forward”. Number of entrances unities are equal than number of emergences unities. The number of emergences unities represents the target of prediction. The prediction “one step before” can be realized with an emergences unities network, and the prediction with “K steps before” can be associate to a network with K emergences unities.
Examples:
1. A model of a neuronal network “one step before”, which use six terms of serial time to predict the next term of series ( 6 neurons at entrance, 15 in hide tier and one in emergence tier, entering = (x…….) , emergence = x….;
2. A model of a neuronal network in to 8 terms of serial time are using for predict the next 2 terms;
3. A model of neuronal network in to 6+2(t-1) terms of serial time are using for predict the next “t” terms.
• the employment of a new index PDG for the prognosis of sweepings administration consulting of:
1. population (the index of variety of population);
2.“the habitat” of sanitation services;
3. economic development:
- the economic development with 2% / year represents an increase of assimilated sweepings amount emanated from shop keeping (Pce – the economic development percent );
- the grows with 3% / year of purchase represent an increase of sweepings (Pv – the purchase grows percent)
4. the amount of sweepings collected from population and economics agents;
5. the amount of sweeping from gardens, plaza and streets;
6. the amount of mud from service station, from buildings and knockings.
For prognosis of sweepings administration development for a certain region and for a certain period of time, are used the scripts. For certain period are constituted the purposes and aims for short , medium and long time. Each script must have the same base which includes the historical evolution of generation system of sweepings for a period between 5 and 7 years. Inside of scripts are constituted the alternatives. Each alternative describe an integrated system of management of sweepings which includes the available and practical operating method of recycle, treatment and elimination of sweepings.
The main criterions of selection for the best script have to accomplish the principles of permanent development:
- to have minimal negative effects about environment;
- to be permissible in socially;
- to be doable in economically;
Keywords:
neural network, sweeping administration, prognosis, economic development, serial time.
SYSTEM MONITORING AND CONTROL USING HIGH COMPLEXITY CONTROL GATES WITH ADVANCED FEATURES
Alexandru GOLOCA1, Vasile Gheorghita GAITAN2, Valentin VLAD3, Valentin POPA4, Vasile BALAU5
Universitatea "Stefan cel Mare" Suceava
1alexg@eed.usv.ro
2gaitan@eed.usv.ro
3vladv@eed.usv.ro
4valentin@eed.usv.ro
Abstract.
In nowadays manufacturing plants there is a certain degree in which the RFID technology has been adopted. This degree may vary from marking the components of a final product (for example a car) using RFID tags in order to verify if they are all in place at the end of the production line, to the production flow monitoring and finally to the supply chain management. The presented system contains advanced High Complexity RFID-based control gates and smart connections to an OPC server that acts as central dispatcher. The system is able to satisfy a large number of customer requirements for process monitoring and control.
The main research project was focused on three types of RFID-based control gates that vary both in price and the package of features that are being offered. The three types are: Low Complexity Control Gates (which will further be referred as LCCGs), Medium Complexity Control Gates (which will further be referred as MCCGs) and High Complexity Control Gates (which will further be referred as HCCGs).
LCCGs are the “low cost” option and they offer: the ability to read fields from RFID tags and save them, the ability to send stored data to a PC, the ability to modify fields on RFID tags as a result of a command from a PC or as the result of an internal script execution, the ability to fire alarms as the consequence of some actions and events that occurred and the ability to control relays and read data using digital inputs and outputs. Further more, this type of RFID gates features a Real Time Operating System, very appropiate for critical situations.
High Complexity Control Gates have been designed as the “top solution”. These offer the same features as the first two types of Control Gates, adding more advanced ones like: an ARM9 microcontroller with MMU support, Arm-Linux operating system and extended web features like an Apache web server optimized for embedded systems and able to run CGI applications as well as user authentication, very large amount of history data (the maximum storage size for Compact Flash Type I is 4GB). The large amount of long term memory allows this kind of Control Gates to offer a great autonomy.
Keywords:
RFID, Linux, Embedded, Process control, OPC.
SPEECH RECOGNITION TECHNIQUES IN COMPUTER ASSISTED LOGOPEDIC THERAPY
Alexandru GOLOCA1, Stefan-Gheorghe PENTIUC2
Universitatea "Stefan cel Mare" Suceava
1alexg@eed.usv.ro
2pentiuc@eed.usv.ro
Abstract.
Nowadays we can find more and more PC applications that offer the possibility to use a relatively new kind of interface, one that eliminates any physical contact: a vocal interface. This kind of applications makes way for a future when human to PC conversation will be just as usual as inter human conversation.
Computer Speech recognition, also known as Automatic Speech Recognition (ASR) consists of a number of software techniques that enable a computer (or even an embedded device) to recognize spoken words in an automated way (without human aid).
This paper presents an ASR designed to offer accurate speech recognition in order to assist a complex monitoring system designed to aid children with speech disabilities.
Speech recognition consists of two problems: isolated word recognition and contextual speech recognition (depending on whether single or multiple words are supposed to be recognized).
There are three main speech recognition techniques: Dynamic Time Warping (DTW, old and obsolete method, it has only an historical importance), Artificial Neural Networks (ANN, they have the ability to learn, just like the human neural networks do, and are very good for small words or phonemes but offer low quality for long words), The Hidden Markov Model (HMM, is the main choice for current ASR systems) and Hybrid models (these use ANNs for phonemes learning and HMMs for deciding the word that generated the given phoneme combination).
The Hidden Markov Model is a statistic model in which the modeled system is assumed to be a “Markov Process” with unknown parameters. The main challenge is to compute the hidden parameters knowing only the visible parameters (those that can be observed). The extracted parameters can be used after that for many different purposes but the main goal is to build a system that is able to classify objects – form the ASR point of view, objects are the spoken words.
The current project follows a number of steps in order to achieve automatic speech recognition. Steps followed at training time are: Sound acquisition, Windowing, Digital filtering, Word Boundary Detection, Feature extraction (using a codebook, which implies algorithms such as K-means and Binary Split), while steps followed at recognition time are: Sound Acquisition, Windowing, Digital Filtering, Word Boundary Detection, Feature Extraction, Distance Computing followed by Recognition.
There is a number of steps that are still under development or experimental: Training a number of HMMs with data from the codebooks and Using the trained HMMs for speech recognition instead of just some simple distance.
Keywords:
Hidden Markov Models, Neural Networks, and Automatic speech recognition.
COMMUNICATION SECURITY IN MOBILE DEVICES USED IN SPEECH THERAPY
Mihai Horia ZAHARIA
Dept. of Computer Science and Engineering
Technical University Gh. Asachi Iasi
Bd-ul D. Mangeron, 53A, 700050, Iasi, Romania
Abstract.
In this paper the problem of securing mobile devices used in speech therapy is presented. The main target consists in making various mobile devices involved in speech therapy to maintain both the confidentiality of personal data of the patient and also to avoid interference when simultaneous communicate with the control center. Due to non-technical type of user all password management will be made automatic by the control system. As result the mobile device will have a user transparent security layer added.
Keywords:
mobile device, distributed computing, communications security
EVALUATION OF RFID LOCATION SYSTEMS
Eugen COCA1, Valentin POPA2
Universitatea "Stefan cel Mare" Suceava
1eugen.coca@usv.ro
2valentin.popa@usv.ro
Abstract.
In this paper we present an evaluation of location systems available commercially on the world market as full working systems or as prototype and evaluation versions. We made tests for both location precision and electromagnetic field emissions in order to find out the possible influences of this kind of systems on human safety.
Radio Frequency Identification (RFID) Systems are based on radio frequency (RF) tags and RF readers. Tags are usually made around a microchip containing a small memory. Modern tags have memory capacities of several tens of kilobits or more. RF readers are microprocessor based systems running software capable of reading and / or writing memory locations from the tags. Multiple working frequencies are used for this kind of systems, from 13.56MHz to 2.4 GHz or even more.
RFID location systems for indoor and outdoor positioning are a promise for the future. The performances of these systems are affected by many factors. We identified here that for a system working in the band near 900 MHz, the objects interposed between the antenna system and the tags to be located may have a great influence in terms of accuracy of the measurement. In closed areas multiple reflection paths may disturb the measurement systems, a percent of only 40 to 60 of total measurements are enough accurate to locate an objects. In such conditions, there are small chances for this kind of systems to be used for high precision applications.
Regarding the EMC aspects of this RFID location system, we may say, based on measurements presented here, that the electric field are high enough not to use this system indoors at distances less than 5 meters, if humans are present on a regular basis in that area. For applications in open areas, like access control for auto vehicles and many similar others, this kind of systems are very good.
Keywords:
location system, RFID, electromagnetic compatibility, emission, spectrum
MATHEMATICAL MORPHOLOGY USED IN IMAGE PROCESSING
Remus Catalin PRODAN1, Valentina PRODAN2
Universitatea "Stefan cel Mare" Suceava
1prodan@eed.usv.ro
2valentina@eed.usv.ro
Abstract.
The article presents the way in which morphologic techniques can be used to process binary images. Starting with the basic operations, erosion and dilatation, defined in mathematical morphology, one entire set of operations can be built in image processing. One of the applications of mathematical morphology is created by filtering the binary images which has a specific sound of “salt-or-pepper”. Meanwhile, the mathematical morphology can be used to eliminate the blanks from the objects present within a scene or to eliminate the objects from a certain aria of the background. The objects touching the edge of the image are not suited for the analysis, therefore, their elimination is required, and in order to do so, a series of morphologic operations can be used. In the conclusion of the article, there is presented a method of segmenting the images in grayscale by using morphologic processing.
Keywords:
image processing, mathematical morphology, hit-and-miss, erosion, dilatation, opening, closing.
STUDY OF ONE DISTRIBUTED DIGITAL SYSTEM FOR MONITORING THE INDUSTRIAL PARAMETERS
Dan MILICI1, Mariana MILICI2, Luciana STAMATE3
Universitatea "Stefan cel Mare" Suceava
1dam@eed.usv.ro
2mami@eed.usv.ro
3lstamate@stud.usv.ro
Abstract.
The data acquisition and transmission system that uses remote serial transmission. The measurement and transmission line has also in its structure the PC compatible computer, the USB cable, the serial cable, the cable through of it is achieved the link between the A52 block and the I-7017 block and, certainly, sensors which measure the desired quantity and which transmit it at I-7017. As we can see, the communication cable in RS-485 protocol can have a large lengthwise till 1,3 km. The communication with the PC can be made either through the I-7560 block (therefore using the USB port of the computer) or connecting directly the A52 block at computer (using the COM port in RS-232 protocol). The acquisition of the desired quantity is made using the LM35 integrated circuit, that is a temperature integrated-sensor. With this kind of system can be acquired until eight different quantities because the data acquisition and transmission block has eight analogical inputs. The acquisition rate is 10 acquires/sec for each analogical input. Due to this fact, the acquisition system can be used for temperature controlling at the starting of one motor, the temperature variation at the testing of one transformer both for the low and the high voltage, overload protections, etc.
The control software is achieved in LabVIEW program. As described previously, this program has one control panel with different control blocks and one diagram panel. The control panel of the application that uses serial protocols is shown in figure 3. The figure has more zones of interest as follows: first zone reprezents the settings field of the serial transmission. We set the baud rate from 1200 bps until 115200 bps, the number of data bits, the number of the parity bits either 0 or 1, or stop bits 0 or 1 and return. The second zone is represented by the settings of the 7000 series’ blocks:
- COM port,
- the block’s address,
- the type of the block,
- if the data verifying through Check SUM is active or nonactive,
- the Time Out settings,
- the number of the channel on which we action and transmit,
- if it must to save the string and the type of data format that has as options 0 or 1.
The serial transmission is recommended to be used in industrial environment due to the fact that the communication distance can vary from a few meters in case of standard serial communication until 1,3 km in case of the RS-485 protocol, with a good immunity at noises and interferences. This immunity confers a large using gamut of the serial transmission, can be used for example to study the temperature increasing of motors’ starting (on electrical driving matter or at varied testing of transformers, high voltage testing, etc.
Keywords:
monitoring, serial interface, parameter, virtual instrument
A MOBILE AND INTELLIGENT DEVICE INTEGRATED IN AN EXPERT SYSTEM FOR PERSONAL LOGOPEDIC THERAPY
Ioan UNGUREAN1, Mihai Gabriel DANILA2, Mihai Gavril ROBU3, Cristian GAITAN4
1,2,3SC GENPRO 07 SRL Suceava
Universitatea "Stefan cel Mare" Suceava
1ioanu32@yahoo.com
2danilamg@eed.usv.ro
3rmgro@yahoo.com
4gaitan@eed.usv.ro
Abstract.
An expert system for personal logopedic therapy must allow the training of children depending on its speech disabilities and former progress. The children’s training is accomplished using exercises chosen by the expert system and can be performed either in the doctor’s office or at home for each child using an intelligent mobile device. The expert system generates a set of exercises for each child depending on the doctor’s recommendation. These exercises are transferred from PC to mobile devices using an Universal Serial Bus connection. The mobile device saves the result of each therapy session and when it is connected to PC it transfers the results to the expert system for analysis. Using the results of these analyses the expert system will decided whether a new session is needed and if that is the case, compute a new set of exercises. In this paper is presented the architecture of mobile device used in logopedic therapy.
Keywords:
expert system, intelligent device, speech terapy
MANUFACTURING WITH INTELLIGENT SHADOW
Leonard IURESCU
Universitatea "Stefan cel Mare" Suceava
iurescu@eed.usv.ro
Abstract.
Nowadays, manufacturing companies are facing with increasing demand of products that must correspond at the variety of consumer’s wishes. So the manufacturing of products must be adaptive and flexible to respond at this variety. Traditional manufacturing, with a hierarchical control system, generate a low-variety of products. Today, as a respond of the new challenge was proposed several approaches generic named intelligent manufacturing systems. One of this is the approach from the view of intelligent agents.
A factory create on the basis of intelligent agent technology is ideal for a good respond to the challenge of the market. But to create such automated factory infrastructure needs substantial investment. So I propose in my work the use of intelligent agents in an old assembly factory (e.g. car assembly factory) with the minimum of investment. This minimum investment in PCs associated to each step of assembly. These PCs are connected in a network. The concept that I propose “manufacturing with intelligent shadow” consist in the fact that for each good that is assembly, a mobile intelligent agent is associated.
When the good is moving in the flow of step assembly the mobile intelligent agent migrate on PCs according to the step. The PCs are the interface between agent (human, machinery, flexible cell, etc) that do the actions to the good and the mobile intelligent agent associated. The mobile intelligent agent knows at start the characteristics that must have the final product and assist the assembly process to materialize them.
So with a minimum investment it possible to convert a traditional factory, with a hierarchical control system, in a intelligent factory with distributed control that allow modification at the level of operation in a short time.
Keywords:
mobile intelligent agent, manufacturing control
MANAGING A PHOTO AND VIDEO DATABASE WITH POSTURES FROM A GESTURE DICTIONARY OF AN INFORMATICS OR ROBOTIC DEVICE WITH GESTURE INTERACTION
Mihai Gabriel DANILA1, Ioan UNGUREAN2, Cristian GAITAN3
1,2SC GENPRO 07 SRL Suceava
3Universitatea "Stefan cel Mare" Suceava
1danilamg@eed.usv.ro
2ioanu32@yahoo.com
3gaitan@eed.usv.ro
Abstract.
Designing and realizing a device with gesture interaction is a very difficult, laborious and complex job. Such as device allows peoples to control an industrial process or a specific system using hands or head gestures, and this kind of commands are useful when controlled system is far away from user or it is in a hardly accessible zone. Current article present a method to manage a distributed database with postures which describe commands possible to be executed by an informatics or robotic device with gesture interaction. Database helps device to learn and test gesture commands. Commands have to be learned in order to be executed whenever the device or the integrated system (a complete system with integrate gesture analyzing and command execution) need that. Our database management system controls photo and video postures existing in database (modify, add, delete, update) and publishes this information on the internet to advantage software detection gestures system to realize learning and testing process with data obtained remotely.
Keywords:
gesture recognition, gesture interaction, distributed database, gesture dictionary, human posture
SPECIFICATIONS FOR DISCOUNTS IN THE INTEGRATED SYSTEM B2B_RFID
Felicia GIZA1, Tudor CERLINCA2, Cristina TURCU3
Universitatea "Stefan cel Mare" Suceava
1felicia@eed.usv.ro
2tudor_c@eed.usv.ro
3cristina@eed.usv.ro
Abstract.
This paper presents the design and development of the discount section for the integrated system RFID_B2B. Because a B2B system treats not only the products of a company, but also his business partners, there are a lot of cases for applying a discount. We analyzed all the possibilities and we can summarize four important cases: one product for a company, all products for a company, one product for all companies and all products for all companies. The paper will present the specification and also the particular cases for each of them. The major problem in specifying and using the discount is the order of applying the discount and also the mode of combining them when there are more than one specified. In this paper, we will also present the implementation of the discount section for the integrated system B2B_RFID. We designed this section in such a way that he can offer a high level of generality. It has two main options: the specifying and the configuring of the discount. The first option permits to specify a discount of one of the four types, depending on the selection made (for example, if one product is selected together with all the companies, than it is considered an 1-N discount etc.). Each type of discount together with his period when is applied is stored in separate tables so they can be used one or another, depending on the user will. If there is more than one type of discount specified for a pair product – company, than the user should follow the configuration option. There he can specify the order of applying the discount and also, if one or more has the same number of order, he can compose or select one of the existing formulas to be applied.
Keywords:
B2B, discount, configuring discount
INTEGRATION OF AN RFID CONTROL GATE IN SUPERVISED PROCESSES
Mihai Gavril ROBU1, Cornel VENTUNEAC2, Cristina GAITAN3, Vasile BALAU4
1,2SC GENPRO 07 SRL Suceava
3Universitatea "Stefan cel Mare" Suceava
4SC CAOM SA Pascani
1rmgro@yahoo.com
2corventu@yahoo.com
3gaitan@eed.usv.ro
4balau@caom.ro
Abstract.
These control gates have been designed as a “compromise solution” between cost and the features that it offers. This device is integrated in distribuited systems; it used to read data stored on a RFID tag using an RFID reader into a data logger, on a Flash memory card (xD – Picture Card) that allows superior storage capacity. The data logger is transferred on request to a PC via RS485 or Ethernet interface. The protocol used is Modbus RS485 or ModBus TCP/IP. The data logger can also be transferred to PC using an FTP connection. There are also available the remote access via a web interface. This allows more remote users to view reports regarding the gate activity viewing a table inside a web page. The information available via the web interface is the number of RFID tags read or the events that occurred during a day.
Keywords:
control gate, RFID Tag, Flash memory, RS485, Ethernet, Modbus, FTP.
PROOF OF CONCEPT OF AN RFID-B2B INTEGRATED APPLICATION
Cristina TURCU1, Tudor CERLINCA2, Remus PRODAN3, Felicia GIZA4, Marius CERLINCA5, Alexandru GOLOCA 6
Universitatea "Stefan cel Mare" Suceava
1cristina@eed.usv.ro
2tudor_c@eed.usv.ro
3prodan@eed.usv.ro
4felicia@eed.usv.ro
5mariusc@eed.usv.ro
6alexg@eed.usv.ro
Abstract.
RFID (Radio-Frequency Identification) technology has been considered as “the next revolution in supply chain” since it allows the tracking of each object or product in real time in the supply chain. This paper examines the impacts and potential benefits generated by a B2B_RFID application in one specific supply chain. A proof of concept is also presented to demonstrate the feasibility of the proposed scenarios. Through a detailed investigation of a business-to-business process, we demonstrate how RFID technology can introduce new opportunities for enriching the enterprise systems with new ways to operate efficiently and better address the demands of business customers. Also, it can be demonstrated that RFID enables more integrated and more collaborative business-to-business (B2B) ecommerce solutions. We can conclude that the paper helps to improve the understanding of the real potential of integrating RFID technologies in B2B applications.
Keywords:
RFID, B2B, tag, supply, chain
INTEGRATING MOBILE APPLICATIONS INTO B2B SYSTEMS
Tudor CERLINCA1, Cristina TURCU2,Felicia GIZA3, Marius CERLINCA4
Universitatea "Stefan cel Mare" Suceava
1tudor_c@eed.usv.ro
2cristina@eed.usv.ro
3felicia@eed.usv.ro
4mariusc@eed.usv.ro
Abstract.
This paper presents the design and development of a mobile application and the way that it was integrated into a more complex RFID_B2B (Business to Business – Radio Frequency IDentification) system. The main goal of such an application is to enlarge the area in which a RFID_B2B system can be used. Usually, the most important requirements for such applications are: to deal with large amount of data, to really work in stand-alone mode and to be easily integrated with complex RFID_B2B systems. We also have focused on the design and development of a new, secure and very fast method for communication and synchronization between: mobile devices and different B2B servers and different mobile devices that run the same application.
Keywords:
B2B, mobile devices, RFID
A N-TIER APPLICATION ARCHITECTURE FOR B2B SYSTEMS
Felicia GIZA1, Tudor CERLINCA2
Universitatea "Stefan cel Mare" Suceava
1felicia@eed.usv.ro
2tudor_c@eed.usv.ro
Abstract.
This paper presents the design and development of a new N-Tier application architecture for B2B systems. Each tier in the architecture will be detailed described from different points of view: purpose, level in the hierarchy and development. We also have focused on the way that we integrated this application into a more complex RFID_B2B (Business to Business – Radio Frequency Identification) system. Mainly, this is the top module of our RFID_B2B system that will store huge amount of data and allow the automatic transaction of goods (that are processed at lower levels) and services between busyness. As the security is playing a very important role in such systems, we did implement secured web services as communication components.
Keywords:
B2B, N-TIER, RFID
PARALLEL COMPUTING SYSTEMS
Carmen PUGHINEANU
Colegiul National “Petru Rares” Piatra Neamt
Str. Stefan cel Mare nr. 4, Piatra Neamt, România
carmen_pughineanu@yahoo.com
Abstract.
This work presents various definitions and ways of parallel computing systems together with its purpose of use and a taxonomy of Flynn and Bell. Further more there are presented ways of programming in parallel computing systems together with the three well-known programming ways in 2 languages and a library of functions, the performance criteria used in designing an interconnecting network and the analysis of their complexity. The purpose of the paper is to show how parallel computing systems work by reducing the execution time.
Keywords:
SISD, SIMD, MISD, MIMD, pipeline, control-flow, massively parallel.
THE USE OF GROUPING TECHNIQUES FOR THE DETERMINATION OF THE HIDDEN LAYERS NUMBER OF A NEURONAL NETWORK
D. OSTAFE1, Stefan-Gheorghe PENTIUC2, Stefan HOLBAN3, E. IANCU4
1,2Universitatea "Stefan cel Mare" Suceava
3Politehnica University of Timisoara
1dumitru@seap.usv.ro
2pentiuc@eed.usv.ro
4eiancu@seap.usv.ro
Abstract.
We consider the set of n patterns for training a neural network, the elements of this set can be grouped, according to a certain criterion, thus obtaining a number of homogenous classes and a number of elements which cannot be associated with any group (unclassified elements). In the case of the specific treatment of the patterns contained in the obtained homogenous class, trough the division into classes of the neural network training patterns set, one can determine the number of hidden layers (the number of groups obtained when the grouping system becomes stable).
Keywords:
neural network, hidden layer, pattern recognition
SYSTEM FOR HAND POSTURES RECOGNITION
Tudor Ioan CERLINCA1, Stefan-Gheorghe PENTIUC2
Universitatea "Stefan cel Mare" Suceava
1tudor_c@eed.usv.ro
2pentiuc@eed.usv.ro
Abstract.
The paper presents a system for recognition of a set of predefined hand posture, aimed to be used in a human-robot interaction. The system is based on a fast and accurate method for hand posture recognition in video sequences using multiple video cameras. The proposed technique is also strong enough in order to deal with changes of lighting and the experimental results shows that it can be used in various vision-based applications that need real-time detection and recognition of hand posture.
The performances of the hand posture classifier were tested in two different environments: static and dynamic. For each method we were interesting to determine the positive and negative detection rates.
For the first one, the performances were tested with sample images that were grabbed from static scenes in which the lighting and the distances from the video cameras were approximately constant. The test set was built using a total of 900 different images (100 images for each posture). For each posture, the positive detection rate that we achieved was in the range of 92-100% and the negative approximately 8%.
For the second method, the performances were tested in real time environments with changes of lighting and distances from the video cameras. The global positive detection rate was in the range of 65-75% and the negative one approximately 15%. The negative detection rate highly depends on the color of the objects in the scene: too many bricks that have a color close to the skin’s one may introduce errors in the detection.
Keywords:
image processing, pattern recognition, hand posture recognition, real-time human-robot interaction