Yazar "Sokullu, Radosveta" seçeneğine göre listele

Listeleniyor 1 - 20 / 48
  • Küçük Resim Yok
    Öğe
    Adaptive packet selection algorithm for bluetooth data packets
    (World Scientific And Engineering Acad And Soc, 2007) Sokullu, Radosveta; Karatepe, Engin; Xu, AP; Zhu, H; Chen, SY; Yan, B; Meng, QG; Miao, D; Fang, Y
    In this paper we present a Matlab-Simulink model for performance evaluation of Bluetooth DHx data packets. The simulation model developed takes into account channel quality, distance and interference in the Bluetooth frequency band. We propose a new adaptive algorithm for packet selection based on the number of retransmissions until successful packet reception under current environment conditions. The results obtained using the simulation model have proved the increase in performance provided by this algorithm.
  • Küçük Resim
    Öğe
    Akıllı ilaç kutusu prototip tasarımı
    (Ege Üniversitesi, 2019) Sokullu, Radosveta; Öner, Betül
    Günümüzde insanların karşılaştıkları çok çeşitli hastalıklar olabilmektedir. Bu hastalıklar insanların hayatını çok farklı şekillerde zorlaştırabilmektedir. Bu hastalıklardan en bilinenlerinden bazıları Alzheimer hastalığı, akıl hastalıkları ve genellikle yaşlılarda bulunan hastalıklardır. Bu hastalıkların tedavisi için düzenli ve doğru dozda ilaç alımı çok önemlidir. Ancak hastalar bazen ilaç almayı unutabilmekte, farklı dozlarda alabilmekte ya da çeşitli sebeplerden ötürü istemeden de olsa ilaçlarını düzenli almayabilmektedir. Bu konudaki çalışmalar farklı sebeplerle olan ilaç alımlarındaki aksamaları en aza indirmenin önemini vurgulamaktadır. Bu projede hastaların alması gereken ilaçları doğru zamanda ve doğru dozlarda almalarının sağlanması amaçlanmıştır. Hastaya yazılan uzun süreli ilaçlar eczacı tarafından, belirlenen dozajlarda sisteme yetkili kişinin tarafında girilecektir. Geliştirilen Akıllı ilaç kutusunun ilk prototipte 3 ilaca kadar ilaç kontrolü yapabilmektedir. Doktor tarafından belirlenen bu ilaçların alınması gereken zamanlarda sisteme eczacı ya da hasta bakıcısı tarafından girilecektir. İlaçları ayarları başka birinin tarafında değiştirmesini parmak izi sensörü ile kontrol edilmekte. Sadece yetkili kişinin parmak izleri ile sistem çalışır. Akıllı ilaç kutusu ilaçların alınması gereken zamanda uyarı (sesli ve görüntülü -ışıklı) yaparak zamanı gelen ilaçları haznesine düşürecektir. Hastanın ilacı alıp almadığı sensör ile kontrol edilecektir. İlaç alınmadığı durumda hastaya uyarı yapılacaktır. Bu uyarı sesli ve ışıklı olduğu için hastanın dikkatini akıllı ilaç kutusuna çekecektir. Haznedeki ilacın alınmaması durumunda hasta yakınlarına ya da hastadan sorumlu kişiye bilgilendirilecektir. Bu projedeki çalışmalar Ege Üniversitesi Araştırma Fonunda (BAP) desteklenmiştir.;Akıllı ilaç kutusu, uyarı sistemi, yaşlılar, ilaç hatırlatma, Alzheimer.;Smart pill box, warning system, elders, medication reminder, Alzheimer.
  • Küçük Resim
    Öğe
    Bluetooth ad hoc uygulamalarının tasarımı
    (Ege Üniversitesi, 2006) Karatepe, Engin; Sokullu, Radosveta
    Bluetooth, performance, adaptive packet selection, chat service, interference.;Bluetooth, performans, adaptif paket seçimi, chat servisi, girişim.;Bu projede gerçekleştirilen çalışmaları dört alt başlıkta toplayabiliriz. Projenin birinci aşamasında Bluetooth ağlarda kullanılan DM1, DM3 ve DM5 data paket standartlarının Matlab-Simulink simülasyon ortamında, ses ve veri iletişimi için yapılmış altyapıyı kullanarak ve IEEE 802.15.1 standardına uygun olarak geliştirdiğimiz DM3 ve DM5 paket tipleri için veri iletişiminin simülasyon modeli sunulmuştur. Bu model ile ortam şartları, mesafe ve eğer varsa diğer Bluetooth modülleri veya genel olarak Bluetooth iletişim frekans aralığına düşen sinyallerin bozucu etkileri dahil edilerek simülasyon ortamında sistem performansları incelenmiştir. Elde edilen sonuçlarda DM5 paket tipi daha fazla kullanıcı data taşıyabilmesine rağmen, gürültülü ortamlarda DM1ve DM3 paket tipleri daha yüksek performans göstermektedir. Projenin ikinci aşamasında Bluetooth teknolojisinde kullanılan DH1, DH3 ve DH5 data paket tipleri için Matlab-Simulink ortamında iletişim performansını değerlendirebilecek bir simülasyon modeli gerçekleştirilmiştir. Geliştirilen bu simülasyon programının doğruluğu literatürde yer alan matematiksel modeller ile kıyaslanarak gösterilmiştir. Ayrıca bu çalışmada literatürdeki çalışmalardan farklı yeni bir adaptif paket seçimi algoritması denenmiştir. Bu algoritmada tekrarlanan frame sayısı kriter olarak seçilmiştir. Elde edilen sonuçlar önerilen algoritmanın performans artışı sağlayabileceğini göstermektedir. Projenin üçüncü aşaması Bluetooth link kalitesi ölçümlerine dayanarak hazırlamış olduğumuz deney düzeneği ile bu deneğin kullanıcı tarafından Bluetooth performansını arttırabilecek algoritmasını ve bu durumu destekleyen sonuçlarını içermektedir. Deneyde dosya transferi uygulamasından faydalanarak Bluetooth?da BER (bit error rate - hata oranı) ve veri aktarım kapasitesi parametrelerinin ortamın link kalitesinden ne derecede etkilendiği araştırıldı. Dosya gönderimi sırasında hem tekli hem de ikili paket secenekleri kullanarak bu paketlerden hangilerinin daha yüksek performansa sahip olduğu gösterilmeye çalışılmıştır. Projenin son aşmasında ise Bluetooth Egitim Araci çatısı altında örnek olarak gerçekleştirilen bir interaktif chat uygulaması anlatılmaktadır. IEEE 802.15 standartlarına uygun Bluetooth teknolojisini temel alarak ve Clarinox firması tarafından ClarinoxBlue algoritması ile Visual C++ programlama dili ile gerçekleştirdiği "Chat Application" programı kullanılmıştır. Bu program üzerinde çeşitli değişiklikler yapılarak ve Net C# programlama dili ile eğitim aracı kapsamında kullanılacak bir arayüz hazırlanmıştır. Bu arayüz üzerinden interaktif örnek olarak "Chat uygulaması" gerçekleştirilmiş ve uygulamanın arka planda nasıl gerçekleştiğini gözlemleyebilmek için çeşitli ekranlar tasarlanmıştır. Bu ekranlar ile "Chat Application" programının çalışması sırasında oluşan ve sürekli değişen trace (iz) dosyaları görsel hale getirilmiştir. Bu dosyalar, program çalışma esnasında bir kere oluşmakta ve programın işleyişine göre sürekli değişmekte ve bize uygulamanın Bluetooth katmanları arasında nasıl gerçekleştiğini göstermektedir.
  • Küçük Resim Yok
    Öğe
    Channel Estimation with Fully Connected Deep Neural Network
    (Springer, 2022) Sokullu, Radosveta; Yildirim, Mete
    In this study, we focus on realizing channel estimation using a fully connected deep neural network. The data aided estimation approach is employed. We assume the transmission channel is Rayleigh and it is constant over the duration of a symbol plus pilot transmission. We develop and tune the deep learning model for various size of pilot data that is known to the receiver and used for channel estimation. The deep learning models are trained on the Rayleigh channel. The performance of the model is discussed for various size of pilot by providing Bit Error Rate of the model. The Bit Error Rate performance of the model is compared to theoretical upper bound which shows that the model successfully estimates the channel.
  • Küçük Resim Yok
    Öğe
    Comparative performance study of ADMR and ODMRP in the context of wireless LANs and Wireless Sensor Networks
    (World Scientific And Engineering Acad And Soc, 2008) Sokullu, Radosveta; Karaca, Ozlem; Demiralp, M; Mikhael, WB; Caballero, AA; Abatzoglou, N; Tabrizi, MN; Leandre, R; GarciaPlanas, MI; Choras, RS
    Wireless LANs (WLANs) and Wireless Sensor Networks (WSNs) are two large groups of networks that have well established application ranges. Despite the fact that they address very distinctive groups of devices and have clearly differentiated wireless interfaces, there are certain similarities which push scientists to look for adopting solutions already designed for WLANs to WSNs. An example of this is the case with routing layer protocols. AODV, an unicast routing protocol, developed for Mobile Ad Hoc networks (MANET), has proved to be applicable and accepted by IEEE as the standard for the routing layer in Low Rate - Wireless Personal Area Networks (LR-WPAN). Still, MANET-originated solutions, including multicast protocols, were initially designed in the context of IEEE 802.11 MAC layer protocol. This paper investigates the feasibility of two popular MANET multicast protocols, ADMR and ODMRP over the IEEE 802.15.4 standard and provides a comprehensive study of the performance of these two protocols with different underlying physical and media access protocols. The protocols have been analyzed with ns-2 network simulator. It appears that even though both protocols are applicable in the selected scenarios, there are specifics in their performance in the context of WSNs which should not be neglected.
  • Küçük Resim Yok
    Öğe
    Comparative Study of Cross Layer Frameworks for Wireless Sensor Networks
    (Ieee, 2009) Karaca, Ozlem; Sokullu, Radosveta
    The totally new communication paradigms introduced by the collaborative sensing notion of wireless sensor networks, materialized by the networked deployment of sensor nodes, has in the recent years forced researchers to re-consider the protocol design process. The traditional layered approach is coming under scrutiny. In its place a new, cross-layer approach has emerged that promises to meet much better the specific needs of versatile applications and severely limited resources characteristic for wireless sensor networks. Lately, the idea of Cross-Layer Frameworks has emerged and the research community is faced with the difficult task to investigate this new platform and its suitability, to weigh its pros and cons in light of its possible future evolution. In this paper we first try to present a brief survey of the suggested so far cross-layer frameworks by investigating the ongoing work in the area of cross-layer frameworks, putting that work in perspective, and consolidating the existing results and insights. Our second aim is to classify the cross-layer frameworks by defining some major criteria and identify their advantages and disadvantages in a unified manner. The work is supported by the evaluation of these frameworks against applications from large international projects.
  • Küçük Resim Yok
    Öğe
    A Comparative Study of MAC protocols for Linear WSNs
    (Elsevier Science Bv, 2015) Sokullu, Radosveta; Demir, Eren; Shakshuki, E
    In this paper, we present an investigation of MAC protocols for linear wireless sensor networks (LWSNs). Recently, there have been a number of applications like roads, bridges and pipeline monitoring, in which the transmission of data over very long lines of consecutive nodes is required. Due to the additional restrictions imposed by the linear topology, such applications present new challenges regarding the provision of low latency and low energy consumption. In order to address these issues researches have proposed a number of protocols tailored specifically for linear topologies. Two different operational mechanisms stand out: the first is based on explicit book and burst transmission, while the second is based on low power listening (LPL) and adaptive sleep cycle. The paper covers the details of the operation of these two types of protocols and provides simulation results for their performance comparison in terms of latency, energy consumption and throughput. The results clearly show the limitations of burst transmission and the advantages of LPL and adaptive sleep cycle as realized in LINE MAC. (C) 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
  • Küçük Resim Yok
    Öğe
    Data Gathering System for Watering and Gas Pipelines Using Wireless Sensor Networks
    (Iaria Xps Press, 2011) Sokullu, Radosveta; Akkas, Mustafa Alper; Demirel, Fahrettin; Mauri, JL; Fries, S; Lopez, MLM; Kovac, RJ
    In this paper we discuss a small size experimentally implemented pipeline watering system, in which the pressure sensors are placed at carefully selected points. In our prototype, the TelosB motes, integrated with the pressure sensors, communicate in real time with each other and also with the base station according to the specifically designed wireless network protocol. The reason to choose a WSN as an infrastructure in our study is to let the system work without the need of extra cabling. Such a system incorporates both efficiency and flexibility, and provides the users with an automatic controlled water/gas system based on the sensor data. In our work, data was continuously measured using a pressure sensor and transferred to a central monitoring station via IEEE 802.15.4 wireless sensor network for storage and display. TelosB wireless motes were programmed with nesC and a graphical user interface was used to capture and display incoming measurements for all selected points being monitored. Evaluation of the system was done based on the WSN performance criteria (packet loss and network lifetime) and also based on the accuracy of the collected pressure data. In both respects the system performed very satisfactory and has been successfully implemented.
  • Küçük Resim Yok
    Öğe
    DEEP LEARNING-BASED DECODING FOR PHASE SHIFT KEYING
    (Union Scientists Bulgaria, 2021) Yildirim, Mete; Sokullu, Radosveta
    In this paper, the application of Deep Learning (DL) in the field of telecommunications is discussed, focusing specifically on symbol detection at the receiver. The performance of Deep Learning-based detection is examined for phase shift keying modulation over Additive White Gaussian Noise (AWGN) and Rayleigh channels. First, a model is proposed which shows that the theoretical bit error rate and throughput can be achieved using DL techniques. Then, the effects of different DL model parameters on the model performance are investigated. The DL model for symbol detection with tuned and minimized parameter set is examined from various aspects and it is shown that this improved version can achieve the desired results with much less complexity of the realization.
  • Küçük Resim
    Öğe
    Demans yaşlılarının çevre destekli yaşamı için akıllı ev yardımcısı
    (Ege Üniversitesi, 2016) Demir, Eren; Akkaş, M. Alper; Sokullu, Radosveta
    İnsanlarda yaşlanmanın bir sonucu olarak zaman içerisinde beyin hücrelerinde hasarlar ve kayıplar olabilmektedir. Bu hasarların bir belirtisi olarak hafıza kayıpları, unutkanlık, dikkat dağınıklığı, yargılama da bozulma, dil ile ilgili sorunlar ve hatta kişilik değişiklikleri meydana gelebilir. Bu belirtileri genel olarak bir arada bulunduran bir hastalık türü olan Demans insanların günlük uğraş ve ilgilerini büyük oranda aksaklığa uğrattığı için son yıllarda bilim insanları tarafından üzerinde çalışılması gereken bir konu haline getirmiştir. Bu konudaki araştırmalar genel olarak yaşlıların yaşadıkları evlerini akıllı eve dönüştürme ihtiyacını ortaya koymuştur. Böylesi bir akıllı evde, ev içerisindeki spesifik konumlara yerleştirilen algılayıcılar yardımıyla anlık olarak gerçekleşen olaylar kayıt altına alınabildiği gibi, evde yaşayan yaşlının konforunu arttırmaya yönelik uyarılar ve bildirimler de yapılabilir. Kablosuz haberleşme teknolojsinin de gelişimiyle birlikte ev içerisinde yerleştirilmiş olan algılayıcılar tarafından üretilen veriler kablosuz olarak merkezi bir veri işleme birimine aktarılabilir hale gelmiştir. Bu noktada kablosuz haberleşme aracı olarak Bluetooth, WiFi veya ZigBee gibi haberleşme protokolleri ve standarları kullanılabilir. Bu protokollerden hangisinin kullanılacağı ise uygulamanın karakteristiğine, ev içi fiziksel koşullara bağımlı olarak değişebilir. Bu projenin amacı Demans hastalığı belirtilerini yoğun bir şekilde taşıyan ve yalnız yaşayan bir hastanın evine yerleştirilen algılayıcılardan kablosuz olarak toplanan ve kronolojik olarak kaydedilen verilerin bir mikrodenetleyici yardımıyla bir bulut sistemine aktarılarak bu verilere internet yardımıyla daha uzak konumlardan dahi ulaşılabilir ve anlık bilgilendirilebilir hale getirmektir. Bu verilerin internet ortamına aktarılması aynı zamanda Nesnelerin İnterneti konseptini de içinde barındırdığından bu çalışma daha farklı bir biçimde önem arz eder konuma gelmiştir. Çalışmada genel olarak kablosuz haberleşme modülü olarak popüler Xbee modülleri ile birlikte mikrodenetleyici olarak Intel Galileo Gen 1 geliştirme platformu kullanılmıştır. Son olarak Galileo merkez birimi tarafından toplanan bu veriler Nesnelerin İnterneti aygıtları için bildirim yapabilme imkanı sunan Pushingbox isimli bulut sistemine aktarılarak, tespit edilen aktivitelerin anlık olarak Android tabanlı bir akıllı telefona iletimi sağlanmıştır.;Demans, Nesnelerin interneti, Intel Galileo, Xbee, Bulut, Pushingbox.;Dementia, Internet of Things, Intel Galileo, Xbee, The Cloud, PushingBox.
  • Küçük Resim Yok
    Öğe
    A Design of Stereoscopic 3D Video Processing System Based on FPGA 3D Formatter in case of FPR
    (Ieee, 2013) Sokullu, Radosveta; Aydin, Mutlu; Yetongnon, K; Dipanda, A; Chbeir, R
    In this paper, we present the details of a video processing system that combines the advantages of Film Patterned Retarder LED/LCD display technology and an FPGA based 3D formatter with enhanced, superior video processing techniques such as formatting, color correction and 3D video depth adjustment. The system includes hardware and software implementation which is designed to process stereoscopic 3D formats using a new methodology. A significant advantage of the system is the possibility to support HDMI 1.4 mandatory formats by using HDMI 1.3 capable System-on-Chip (SoC) scaler IC. The 3D function is provided by FPGA based 3D formatter. Because of the reconfigurable enhanced video blocks, the suggested FPGA based formatter can be designed as an ASIC chip to be used in TVs and set-top box (STB) applications.
  • Küçük Resim Yok
    Öğe
    Effects of LTE Random Access Methods on Healthcare M2M Applications
    (Elsevier Science Bv, 2017) Balci, Abdullah; Sokullu, Radosveta; Shakshuki, E
    In recent years, the communication between machines, known as Machine-to-Machine (M2M) communication, has become a dynamic research area. M2M refers to networks, which connect devices together and enable them to make a smart decision. The most viable option of implementing M2M communication is over the Long Term Evolution (LTE) network. A key challenge for M2M communication over the Long Term Evolution (LTE) is the performance of Random Access Channel (RACH) because its operation was not designed with the enormous number of M2M devices in mind. To deal with the situation reduces network access success rate, several major access control methods were adopted. However, the suggested methods generally increase the access success rate regardless of delay requirements. On the other hand, access delay is crucial for many M2M applications such as healthcare. This paper investigates the effects of LTE Random Access Methods on healthcare M2M applications when enormous number of devices are introduced. The study clearly shows that to meet the requirements for delay sensitive M2M applications, the random access (RA) resources should be very strictly and carefully allocated. (c) 2017 The Authors. Published by Elsevier B.V.
  • Küçük Resim Yok
    Öğe
    Enhancing Performance of M2M Random Access in 3GPP LTE Networks
    (Ieee, 2017) Balci, Abdullah; Sokullu, Radosveta; Akkas, Mustafa Alper
    The proliferation of smart devices that can autonomously connect and exchange information over the internet has given rise to a new dynamic area of research, Machine-to-Machine (M2M) communications. M2M refers to networks, which connect devices together and enable them to make smart decisions based on the generated and/or transferred data. The most viable option of implementing M2M communications is over cellular networks. These networks are designed for Human-to-Human (H2H) communication, which is characterized by smaller number of connection requests and longer connection times as compared to the thousands of M2M that will try to connect simultaneously to transmit only a minimum amount of data. When such enormous number of devices is introduced, the existing network will be exposed to high collision rate and extreme connection delays. In LTE, the uplink resources for random access (RA) procedure will be insufficient, and the number of successfully connected devices will bottom out. 3GPP has proposed some solutions to overcome the congestion during the random access procedure. Access Class Barring (ACB) is the most efficient method which increases the access success rate, but it has a disadvantage in terms of access delay. In this paper, we propose to control the ACB parameters adaptively in order to maximize the use of RA resources. Different from other existing methods, our method is based on predicting the number of accessing devices at the beginning of each RA slot and then adjusting the ACB parameters accordingly. Our results show that, the proposed scheme improves delay up to nearly 70% and converges to optimal throughput.
  • Küçük Resim Yok
    Öğe
    The European Network of Excellence CRUISE application framework and network architecture for wireless sensor networks
    (Ieee, 2007) Murray, Ken; Timm-Giel, Andreas; Becker, Markus; Guo, Cheng; Sokullu, Radosveta; Marandin, Dimitri
    This paper presents the main application scenarios under investigation within the Network of Excellence CRUISE, which is funded in the IST 6th Framework Programme. First an application scenario framework is presented being used for the design process of wireless sensor networks (WSN). Using the framework WSN applications can he compared and contrasted in terms of user, systems requirements, and object interaction. Following the applications are mapped to the sensor network architecture applied in CRUISE and described in the paper. Finally the application scenarios are mapped into the network architecture to demonstrate its generic structure and applicability across multiple application domains.
  • Küçük Resim Yok
    Öğe
    Face Recognition Based Multifunction Smart Mailbox
    (Ieee, 2020) Sokullu, Radosveta; Akin, Ozkan; Aslan, Erginay
    The abundance and variety of the Internet of Things (IoT) applications in the recent couple of years has reach a new record. From smart home, smart traffic, smart city, to industry, logistics and agriculture IoT applications are penetrating in every facet of our life and our society. A lot of simple and more complicated solutions benefit from the flexibility and diversity of connections that this new technology provides. In order to provide better management, better visualization, increase scale and reduce response times logistics and warehouse management are embracing new ways to improve upon the already old RFID technology. In this sense IoT provides an unprecedented technology to meet all these requirements in a uniform and easy to manage way. In this work a specific smart application - IoT mailbox with face recognition is discussed. Combining cellular connectivity with image processing it ensures the user that his valuable documents will be securely delivered. The developed prototype consists of a fingerprint reader, a camera, electromagnetic lock, a small LCD screen, a microphone and a loudspeaker all connected to an Arduino Uno which processes the data and establishes the network connectivity through a GSM module. Additionally, to perform face detection and recognition based on a prerecorded image set the camera is connected to Raspberry Pi and OpenCV and Python software is employed. The remotely controlled electromagnetic lock ensures that shipments are kept protected until the designated receiver takes possession of them.
  • Küçük Resim Yok
    Öğe
    Fairness aware deep reinforcement learning for grant-free NOMA-IoT networks
    (Elsevier, 2024) Balci, Abdullah; Sokullu, Radosveta
    Next generation networks have special areas related with the Internet of Things (IoT) to improve the performance of cellular networks in terms of throughput. Grant -free non -orthogonal multiple access (GF-NOMA) seems a feasible solution, letting machine type communication (MTC) devices transmits their packets when they ready to transmit. GF-NOMA increases the spectral efficiency by using the superimposing signals with different power levels over the same time and frequency resources. However, the main drawbacks of GF-NOMA are randomness and the management of power level selection of MTC devices. In 6G-IoT networks, the intelligence should be met to random access. It is time to design new access methods to solve the GFNOMA issues that should be between the randomness and fully coordinated medium access. Deep -Q -Network (DQN) has become a very hot research topic in recent years that let the MTC devices to make a smart decision in an intelligent way to improve the throughput. Selfishness is an undesirable behavior of DQN for GF-NOMA system where the resources have different cost. In this study, we develop a novel learning framework for power domain GF-NOMA. The goal of our learning framework is to maximize the throughput considering fairness in power consumption which provides long -life to the IoT network. The learning algorithm push the MTC devices to exchange the resources between each other over time. The results show that the proposed method outperform the NOMA scheme with random selection in terms of throughput and increase the fairness index when the DQN with selfish behavior is employed.
  • Küçük Resim Yok
    Öğe
    Fairness aware deep reinforcement learning for grant-free NOMA-IoT networks (vol 25, pg 1, 2024)
    (Elsevier, 2024) Balci, Abdullah; Sokullu, Radosveta
    [Abstarct Not Available]
  • Küçük Resim Yok
    Öğe
    Fault Management for Smart Wireless Sensor Networks
    (Ieee Computer Soc, 2012) Sokullu, Radosveta; Karaca, Ozlem; Apduhan, BO; Hsu, CH; Dohi, T; Ishida, K; Yang, LT; Ma, J
    Smart environments are discussed as the next step in intelligent systems automation related to utilities, construction, industry, home and transportation. In the core of these environments are the wireless sensor and actuator networks (WSAN) which have seen an unprecedented development in the last decade both in terms of complexity and supporting concepts [1]. In this respect, there are still a number of issues that have to be solved in new ways suitable to the specifics and requirements of the smart environment. One such pressing issue is providing high level of fault tolerance. Fault tolerance is defined as the ability of a network to continue functioning in the presence of link and node failures. As the wireless media has highly varying characteristics and the nodes are easily prone to failures, the development of new fault tolerance algorithms is a pressing question. Furthermore, traditional layered protocol design has proved ineffective for WSAN and the new cross layer design approach has emerged as a foundation for protocol design in smart environments. The paper presents integrated fault tolerance module for WSAN based on the cross layer design approach. The results presented, based on packet loss, delay and energy consumption, support the thesis that adoption of the suggested fault management module improves the general performance of the network.
  • Küçük Resim Yok
    Öğe
    Fault Tolerance and Fault Management Issues in Wireless Sensor Networks: Recent Developments
    (Int Frequency Sensor Assoc-Ifsa, 2015) Sokullu, Radosveta; Akkan, Ozlem Karaca; Yurish, SY; Malayeri, AD
  • Küçük Resim Yok
    Öğe
    Frame Detection with Deep Learning
    (2021) Yıldırım, Mete; Sokullu, Radosveta
    Deep learning has become a way of solution for the realization of complex computations. As electronic communication starts to use more complex channels, the systems need to handle tough computations. For this reason, research on the use of deep learning in communication has increased recently. These researches aim to realize many applications used in communication with deep learning. Frame detection is one of the first things a receiver must handle and it may require a lot of hard computations. Deep learning-based frame detection can be an alternative approach. This study aims to build models that perform frame detection with deep learning. The proposed models provide the performance of correlation-based frame receivers commonly used for frame detection. The mean square root error of the prediction deviation is used as an evaluation metric to compare the proposed model to classic systems.