5916 research outputs found
Sort by
Determining the structural life of an electric vehicle battery from road loads
No full textAlthough the electrical system is important in the development of electric vehicles, the durability of the battery system against mechanical loads and the performance of the battery against shock reactions are just as important. The objective of this study is to develop a methodology to determine the structural life of the battery of the C-platform sports utility vehicle (SUV). For this purpose, a power spectral density (PSD) durability test profile is generated and compared to battery test standards such as ISO 6469:2019, AK-LH 5.21 and SAE J2380. Analytical Virtual Proving Ground (VPG), a multibody dynamics simulation model, is also developed and correlated to test data. The results show that FDS values for AK-LH is higher than the fatigue damage of the collected vehicle data, while the FDS results for ISO standard are lower compared to the vehicle data. The results also indicate that the loads in the longitudinal (x-direction) and lateral (y-direction) directions are different. Therefore, loads of different amplitudes should be used for these directions, contrary to SAE J2380 and USABC standards. Finally, it is concluded that the VPG model can be used for determining the fatigue life when there is no test data, thanks to its high accuracy
Görünür ışık iletişimi için uyarlanabilir çoklu girişli çoklu çıkış (MIMO) teknikleri
No full textVisible Light Communication (VLC) technology is a promising solution of wireless communication systems in beyond 5G communication networks, which provides low implementation cost, high energy efficiency, and high-speed data transmission. The VLC systems can be considered as a complementary or alternative for the existing Radio Frequency (RF) based wireless communication systems to satisfy the demands of increasing high capacity. In the VLC systems, Light-Emitting Diodes (LEDs) are used as transmitters, and Photodetectors (PDs) are used as receivers. Multiple LEDs are installed to offer sufficient illumination for indoor environments. This feature can be utilized to implement Multiple-Input Multiple-Output (MIMO) communications systems in order to achieve high data rates by improving spectral efficiency. However, LEDs can only convey unipolar signals. Furthermore, the nature of the VLC channel is frequency-selective, which causes significant degradation in the performance of the VLC system due to critical Inter-Symbol Interference (ISI). Several schemes based on pulse modulation and Orthogonal Frequency Division Multiplexing (OFDM) have been proposed to overcome the ISI of the VLC system. A new Optical Orthogonal Frequency Division Multiplexing (O-OFDM), referred to as a Unipolar Orthogonal Frequency Division Multiplexing (U-OFDM), appears as an attractive solution for emerging VLC systems. In the future of 5G and beyond 5G communication networks, the spectral efficiency needs further improvement through MIMO with adaptive transmission technique. Motivated by these, we focus on improving the spectral efficiency and Bit Error Rate (BER) performance for the MIMO VLC systems in both the single-carrier and multi-carrier. Firstly, we propose a novel Adaptive Spatial Modulation (ASM) scheme, referred to as Flexible Generalized Spatial Modulation (FGSM), for single-carrier MIMO VLC systems to achieve better average Symbol Error Rate (SER) and higher spectral efficiency with a fixed overall number of LEDs compared to existing ASM and Generalized Spatial Modulation (GSM) schemes. The proposed system varies the modulation sizes over the available LEDs as well as the number of active LEDs. The selection of the modulation sizes is based on an optimization problem for the average SER under a predefined value of the spectral efficiency. We derive a closed-form expression of approximate average SER for the proposed system and evaluate its decoding complexity. We show that the FGSM scheme can be a potential candidate for future MIMO-VLC systems. Secondly, we firstly propose adaptive transmission technique for the MIMO VLC system in conjunction with U-OFDM to exploit the U-OFDM benefits. The proposed adaptive MIMO U-OFDM VLC system is implemented to support three different MIMO modes that enable a set of different modulation sizes. The considered MIMO modes are Repetition Coding (RC), Spatial Multiplexing (SMUX) and Spatial Modulation (SM). In the RC, the same signal is transmitted simultaneously from all LEDs, while independent signals are transmitted simultaneously from all LEDs in the SMUX. In the SM technique, a single LED is activated and the index of the activated LED carries information in addition to modulated signals. Depending on the received Signal-to-Noise Ratio (SNR) and target BER, the proposed adaptive transmission system switches between the available MIMO modes and adjusts its modulation size to achieve higher spectral efficiency. The proposed U-OFDM system is applied over different VLC MIMO setups with realistic channel models for 8×8, 4×4 and 2×2 MIMO systems. Recently, Generalized LED Index Modulation (GLIM) appears as an attractive MIMO OFDM scheme for emerging VLC systems. Therefore, in the last part of this thesis, we propose a Magnitude and Wrap-Phase OFDM (MW-OFDM) scheme for the MIMO VLC systems. The proposed MW-OFDM scheme relies on the conversion of complex signals into polar form with the magnitudes and wrap-phases to decrease the restriction on the number of LEDs compared to the conventional GLIM-OFDM. Moreover, the Maximum Likelihood (ML) estimator for the proposed scheme is derived. The proposed MW-OFDM scheme improves the average bit error rate and provides a significant reduction in the decoding complexity, compared to the conventional GLIM-OFDM. Moreover, a half number of LEDs are required for the proposed scheme to deliver the same spectral efficiency in a comparison with the conventional GLIM-OFDM.Görünür Işık Haberleşme (VLC) teknolojisi; 5G'nin ötesindeki ağlarda kablosuz iletişim sistemlerinin umut verici bir çözümüdür. Düşük uygulama maliyeti, yüksek enerji verimliliği ve yüksek hızla veri iletimi sağlar. VLC sistemleri, yüksek kapasitenin artırılması taleplerini karşılamak için mevcut Radyo Frekansı (RF) tabanlı kablosuz iletişim sistemleri için tamamlayıcı veya alternatif olarak düşünülebilir. VLC sistemlerinde, ışık yayan Diyotlar (LED'ler) geri-verici olarak ve Fotodetektörler (PD'ler) alıcı olarak kullanılır. İç mekan ortamları için yeterli aydınlatma sağlamak üzere birden fazla LED monte edilmiştir. Bu özellik, spektral verimliliği artırarak yüksek veri hızlarına ulaşmak için Çok Girişli Çok Çıkışlı (MIMO) iletişim sistemlerini uygulamak için kullanılabilir. Ancak, LED'ler yalnızca tek kutuplu sinyalleri iletebilir. Ayrıca, VLC kanalının doğası frekans seçicidir, bu da kritik Semboller Arası Girişim (ISI) nedeniyle VLC sisteminin performansında önemli ölçüde bozulmaya neden olur. VLC sisteminin ISI'sinin üstesinden gelmek için darbe modülasyonu ve Ortogonal Frekans Bölmeli Çoğullamaya (OFDM) dayanan birkaç şema önerilmiştir. Yeni Optik Ortogonal Frekans Bölmeli Çoğullama (O-OFDM), Tek Kutuplu Ortogonal Frekans Bölmeli Çoğullama (U-OFDM) olarak adlandırılan, ortaya çıkan VLC sistemleri için çekici bir çözüm olarak görülmektedir. 5G'nin geleceğinde ve 5G iletişim ağlarının ötesinde, spektral verimliliğin adaptif iletim tekniği ile MIMO aracılığıyla daha da geliştirilmesi gerekmektedir. Bunları motive ederek, öncelikle MIMO VLC sistemlerinin daha iyi ortalama Symbol Error rate (SER) ve daha yüksek spektral verimlilik elde etmek için sabit bir toplam sayıda Mevcut ASM ve Generalized Spatial Modulation (GSM) şemalarıyla karşılaştırıldığında LED'ler. Önerilen sistem, mevcut LED'lerin yanı sıra aktif LED'lerin sayısına göre modül-asyon boyutlarını değiştirir. Modülasyon boyutlarının seçimi, spektral verimliliğin önceden tanı-mlanmış bir değeri altında ortalama SER için bir optimizasyon problemine dayanmaktadır. Önerilen sistem için yaklaşık ortalama SER'nin kapalı formda bir ifadesini türetiriz ve kod çözme karmaşıklığını değerlendiririz. FGSM şemasının gelecekteki MIMO-VLC sistemleri için potansiyel bir aday olabileceğini gösteriyoruz. İkinci olarak, U-OFDM avantajlarından yararlanmak için MIMO VLC sistemi için U-OFDM ile birlikte uyarlamalı iletim tekniği öneriyoruz. Önerilen uyarlanabilir MIMO U-OFDM VLC sistemi, bir dizi farklı modülasyon boyutu sağlayan üç farklı MIMO modunu desteklemek için uygulanır. Dikkate alınan MIMO modları Tekrar Kodlama (RC), Mekansal Çoklama (SMUX) ve Mekansal Modülasyon (SM) 'dir. RC'de, aynı sinyal tüm LED'lerden eşzamanlı olarak iletilirken, bağımsız sinyaller SMUX'daki tüm LED'lerden aynı anda iletilir. SM tekniğinde, tek bir LED etkinleştirilir ve etkinleştir-ilen LED'in dizini modüle edilmiş sinyallerin yanı sıra bilgi taşır. Alınan Sinyal-Gürültü Oranına (SNR) ve hedef Bit Hata Oranına (BER) bağlı olarak, önerilen uyarlamalı iletim sistemi mevcut MIMO modları arasında geçiş yapar ve daha yüksek spektral verimlilik elde etmek için modülasyon boyutunu ayarlar. Önerilen U-OFDM sistemi, 8x8, 4x4 ve 2x2 MIMO sistemleri için gerçekçi kanal modellerine sahip farklı VLC MIMO kurulumlarına uygulanır. Son zamanlarda, Generalized LED Index Modulation (GLIM), ortaya çıkan VLC sistemleri için çekici bir MIMO OFDM şeması olarak ortaya çıkmaktadır. Bu nedenle, bu tezin son bölümünde, MIMO VLC sistemleri için bir Magnitude and Wrap-Phase OFDM (MW-OFDM) şeması öneriyoruz. Önerilen MW-OFDM sistemi, geleneksel GLIM-OFDM'e kıyasla LED sayısındaki kısıtlamayı azaltmak için karmaşık sinyalin büyüklüğe ve sargı fazına dönüştürülmsini kullanır. Ayrıca, önerilen şema için bir ML Tahmincisi (MLE) türetilir. Geleneksel GLIM ile karşılaştırıldığında, önerilen MW-OFDM aynı spektral verimliliği sağlamak ve aynı zamanda ortalama BER'i geliştirmek için yarım sayıda LED gerektirir. Ayrıca, önerilen MW-OFDM'in kod çözme karmaşıklığı geleneksel GLIM-OFDM ile kıyaslandığında önemli ölçüde azalır
Numerical analysis of the radiant heating effectiveness of a continuous glass annealing furnace
No full textRadiative heat transfer is the dominant mode of heat transfer in glass annealing furnaces. Especially in electrically heated furnaces due to the nonparticipating nature of the furnace atmosphere the radiative heat transfer predominantly occurs between surfaces. In this regard, in container glass manufacturing the layout of the container glasses inside the furnace can considerably affect the radiative heating effectiveness of the furnace. In this numerical study, the effect of the spacing between the bottle rows on the radiant heating effectiveness of the furnace was numerically investigated. An industrial scale continuous annealing furnace model was used and the combined conduction – radiation heat transfer modes in the furnace were solved with the in-house developed transient solver. The convective heat transfer inside the furnace was not taken into account due to its relatively lower contribution in heat transfer. The computational cost of the numerical model is reduced with the use of a computational domain consisting of 7 bottle rows instead of full 29 bottle rows. Reverse Monte Carlo Ray Tracing method is used to solve for the surface-to-surface radiative heat transfer inside the radiant heating zone and integrated into GPU to reduce its computational cost. Further computational speed-up was achieved with the use of a 2nd order accurate radiation boundary condition implementation. It allows a time step increase of 2.5x in comparison with the first order accurate implementation for the same level of accuracy. Finally, a parametric study on the effect of the spacing between the bottle rows on the radiative heating of the bottles have been conducted. In this regard, the spacing between the bottle rows and the conveyor speed have been altered in a way that the throughput of the furnace, the number bottles processed per unit time, is kept constant. The study shows that there is an optimum spacing value between the bottle rows. For the particular bottle geometry considered in this study, the optimum row spacing value is found to be 35% of the bottle diameter. Running the furnace at this optimum value, the length of the heating zone could be reduced by 8.5% compared to the case when the furnace was operated with the bottles almost in touch with each other.TÜBİTA
Imitation and mirror systems in robots through Deep Modality Blending Networks
No full textLearning to interact with the environment not only empowers the agent with manipulation capability but also generates information to facilitate building of action understanding and imitation capabilities. This seems to be a strategy adopted by biological systems, in particular primates, as evidenced by the existence of mirror neurons that seem to be involved in multi-modal action understanding. How to benefit from the interaction experience of the robots to enable understanding actions and goals of other agents is still a challenging question. In this study, we propose a novel method, deep modality blending networks (DMBN), that creates a common latent space from multi-modal experience of a robot by blending multi-modal signals with a stochastic weighting mechanism. We show for the first time that deep learning, when combined with a novel modality blending scheme, can facilitate action recognition and produce structures to sustain anatomical and effect-based imitation capabilities. Our proposed system, which is based on conditional neural processes, can be conditioned on any desired sensory/motor value at any time step, and can generate a complete multi-modal trajectory consistent with the desired conditioning in one-shot by querying the network for all the sampled time points in parallel avoiding the accumulation of prediction errors. Based on simulation experiments with an arm-gripper robot and an RGB camera, we showed that DMBN could make accurate predictions about any missing modality (camera or joint angles) given the available ones outperforming recent multimodal variational autoencoder models in terms of long-horizon high-dimensional trajectory predictions. We further showed that given desired images from different perspectives, i.e. images generated by the observation of other robots placed on different sides of the table, our system could generate image and joint angle sequences that correspond to either anatomical or effect-based imitation behavior. To achieve this mirror-like behavior, our system does not perform a pixel-based template matching but rather benefits from and relies on the common latent space constructed by using both joint and image modalities, as shown by additional experiments. Moreover, we showed that mirror learning (in our system) does not only depend on visual experience and cannot be achieved without proprioceptive experience. Our experiments showed that out of ten training scenarios with different initial configurations, the proposed DMBN model could achieve mirror learning in all of the cases where the model that only uses visual information failed in half of them. Overall, the proposed DMBN architecture not only serves as a computational model for sustaining mirror neuron-like capabilities, but also stands as a powerful machine learning architecture for high-dimensional multi-modal temporal data with robust retrieval capabilities operating with partial information in one or multiple modalities.European Union’s Horizon 2020 ; Japan Science and Technology Agency, Japan CREST "Cognitive Mirroring" ; International Joint Research Promotion Program of Osaka University, Japan un-der the project "Developmentally and biologically realistic mod-eling of perspective invariant action understanding" ; Turkish Directorate of Strategy and Budget under the TAMPublisher versio
Robust parameter design and optimization for quality engineering
No full textThis paper proposes a methodology to determine the optimal settings of key decision variables that affect the resilience of an engineering design against uncertainty. Uncertainty in quality engineering is often caused by environmental factors, and scarcity of data due to limitations in the experimentation phase amplifies the level of ambiguity. The proposed robust parameter design and optimization approach utilizes the Taguchi method to find critical variables to be used in the optimization, and it utilizes robust optimization to immunize the obtained solution against uncertainty. To demonstrate our approach, we focus on design optimization of an injection molding product, a refrigerator door cap, made from thermoplastic raw material and its key quality characteristic, warpage. The near-optimal designs found by the robust parameter design and optimization approach are implemented in a real-life manufacturing environment. The numerical experiments show that the new designs significantly improve the warpage quality characteristic and the total production cycle time compared to the current design used in the manufacturing company.Home Appliances Department of the Vestel Electronics Compan
İki taraflı müzakere için aktör-kritik pekiştirmeli öğrenme yaklaşımı
No full textDesigning an effective and intelligent bidding strategy is one of the most compelling research challenges in automated negotiation, where software agents negotiate with each other to find a mutual agreement when there is a conflict of interests. Instead of designing a hand-crafted decision-making module, this thesis proposes a novel bidding strategy adopting an actor-critic reinforcement learning approach, which learns what to offer in a bilateral negotiation. An entropy reinforcement learning framework called \acrfull{sac} is applied to the bidding problem, and a self-play approach is employed to train the model determining the target utility of the coming offer based on previous offer exchanges and remaining time. Furthermore, an imitation learning approach called behavior cloning is adopted to speed up the learning process. Also, a novel reward function is introduced that does not only take the agent's own utility, but also the opponent's utility at the end of the negotiation. The developed agent is empirically evaluated. Thus, a large number of negotiation sessions are run against a variety of opponents selected in different domains varying in size and opposition. The agent's performance is compared with its opponents and the performance of the baseline agents negotiating with the same opponents. The empirical results show that our agent successfully negotiates against challenging opponents in different negotiation scenarios without requiring any former information about the opponent or domain in advance. Furthermore, it achieves better results than the baseline agents regarding the received utility at the end of the successful negotiations.Bir çıkar çatışması olduğunda yazılım aracılarının karşılıklı bir anlaşma bulmak amacıyla birbirleriyle yaptığı pazarlık için etkili ve akıllı bir teklif stratejisi tasarlamak otomatik müzakeredeki en zorlayıcı araştırma zorluklarından biridir. Bu tez de, elle hazırlanmış bir karar verme modülü tasarlamak yerine, ikili bir müzakerede hangi teklifi sunulacağını öğrenen bir aktör-eleştirmen takviyeli öğrenme yaklaşımını benimseyen yeni bir teklif stratejisi önerilmiştir. Teklif verme yaklaşımı için Soft Actor-Critic (SAC) adı verilen bir entropi pekiştirmeli öğrenme yaklaşımı kullanılmıştır. Önceki teklif değişimlerine ve kalan süreye dayalı olarak gelecek teklifin hedef değerini belirleyen SAC modeli eğitmek için bir kendi kendine oynama yaklaşımı kullanılır. Ayrıca, öğrenme sürecini hızlandırmak için davranış klonlama adı verilen bir taklit öğrenme yaklaşımı benimsenmiştir. Bunlara ek olarak, müzakerenin sonunda yalnızca temsilcinin aldığı teklif değerini değil, aynı zamanda rakibin teklif değerinide kullanan yeni bir ödül işlevi tanıtıldı. Bu tez kapsamında geliştirilen SAC ajanı ampirik olarak değerlendirildi. Bu amaçla, büyüklük ve zorluk bakımından farklı alanlarda seçilen çeşitli muhaliflere karşı çok sayıda pazarlık oturumları yürütüldü. Ajanın performansı rakipleriyle ve aynı rakiplerle pazarlık eden temel ajanların performansıyla karşılaştırıldı. Ampirik sonuçlar, temsilcimizin, rakip veya alan hakkında önceden herhangi bir bilgi gerektirmeden farklı pazarlık senaryolarında zorlu rakiplere karşı başarılı bir şekilde pazarlık ettiğini göstermektedir. Ayrıca, başarılı pazarlık sonunda alınan teklif değeri göz önünde bulundurulduğunda temel aracılardan daha iyi sonuçlar elde etmektedir
Cyclical inversion of limits and centres: the formation process of the Regio quartadecima, Constantinople
No full textThe paper reconstructs the topography of Constantinople’s fourteenth region (regio XIV) applying the urban morphology
analysis methods (Caniggia and Maffei, 1979) and the attractors’ theory (Camiz, 2018) to the fragmentary documental
sources and scarce archaeological data. The pontem sublicium sive ligneum’s location was determined as part of a street
network, in analogy with the pons sublicius in Rome, according to the formation process of the territorial organism. This
was the starting point for the reconstruction of the topographic mosaic. By redefining the path of the Constantinian walls
upon quantitative sources it was possible to localise the monumental buildings of the XIV region, as listed in the Notitia
Urbis Constantinopolitanae, with reference to the morphology of the territory described by Dionysius of Byzantium and
the Patria Konstantinopoleos. The form of the territory is a permanent element within urban contexts of continuous
changes, demolitions and reconstructions. The analysis of the urban tissues, the road network’s diachronic attraction
and the reconstruction of the territorial organism provided the general methodological framework for the placement of
the topographical urban fragments mentioned by historical sources upon a GIS
Landscape characteristics and main plant species of floristic composition of Doğanbey old village in Aydın
No full textDoganbey old village is a historic rural settlement that is located in the national park of Dilek Peninsula Buyuk Menderes Delta. The village reflects rural and unique landscape characteristics and units with both natural and human-influenced elements. The aim of this study is to determine floristic composition of the village and to confirm landscape characteristics by field study. As the methodology; empirical analyses were implemented to determine rural landscape characteristic units and plant species identification at the site. Both native and cultural plant species of the village were determined at the site. Results support the uniqueness of the village with its endemic plant species and its rural habitat. In conclusion; touristic potential of the village should be highlighted with a pro- environmental approach, and future implication or development plans should consider biodiversity conservation from the aspect of sustainability of the village.Publisher versio
A data-driven matching algorithm for ride pooling problem
No full textThis paper proposes a data-driven matching algorithm for the problem of ride pooling, which is a transportation mode enabling people to share a vehicle for a trip. The problem is considered as a variant of matching problem, since it aims to find a matching between drivers and riders. Proposed algorithm is a machine learning algorithm based on rank aggregation idea, where every feature in a multi-feature dataset provides a ranking of candidate drivers and weight for each feature is learned from past data through an optimization model. Once weight learning and candidate ranking problems are considered simultaneously, resulting optimization model becomes a nonlinear bilevel optimization model, which is reformulated as a single level mixed-integer nonlinear optimization model. To demonstrate the performance of the proposed algorithm, a real-life dataset from a mobile application of a ride pooling start-up company is used and company's current approach is considered as benchmark. Results reveal that proposed algorithm correctly predicts the first choice of riders 17% to 28% better compared to the benchmark in different scenarios. Similarly, proposed algorithm offers recommendation lists in which the preferred driver is ranked 0.38 to 1.12 person closer (to the rider's actual choice) compared to the benchmark
A computer science-oriented approach to introduce quantum computing to a new audience
No full textContribution: In this study, an alternative educational approach for introducing quantum computing to a wider audience is highlighted. The proposed methodology considers quantum computing as a generalized probability theory rather than a field emanating from physics and utilizes quantum programming as an educational tool to reinforce the learning process. Background: Quantum computing is a topic mainly rooted in physics, and it has been gaining rapid popularity in recent years. A need for extending the educational reach to groups outside of physics has also been becoming a necessity. Intended Outcomes: This study aims to inform academics and organizations interested in introducing quantum computing to a diverse group of participants on an educational approach. It is intended that the proposed methodology would facilitate people from diverse backgrounds to enter the field. Application Design: The introductory quantum physics content is bypassed and the quantum computing concepts are introduced through linear algebra instead. Quantum programming tasks are prepared in line with the content. Pre/post-test design method and Likert scale satisfaction surveys are utilized to measure knowledge acquisition and to evaluate the perception of the learning process by the participants. Findings: Conducted pre/post-test design survey shows that there is a statistically significant increase in the basic knowledge levels of the participants on quantum computing concepts. Furthermore, no significant difference in the gain scores is observed between the participants from different STEM-related educational backgrounds. The majority of the participants were satisfied and provided positive feedback