Fusaomi NagataProfessor, Department of Mechanical Engineering, Faculty of Engineering, Sanyo-Onoda City University, Japan
Speech Title: Transfer Learning Based CNN and Visual Feedback Control for a Pick and Place Robot
Abstract: Artificial neural network (ANN) which has five or more layers structure is called deep NN (DNN) and it has been recognized as one of the most powerful machine learning techniques. Convolutional neural network (CNN) has a reasonable structure for image recognition, so that it has been being applied to defect inspection processes in various industrial manufacturing lines. It is also known that support vector machine (SVM) has a superior ability for binary classification in spite of only having two layers. The authors have already developed a CNN&SVM design and training tool for easy consideration of defect detection systems, while the effectiveness and the validity have been proved through several CNNs design, training and evaluation [1, 2]. The tool further enables to facilitate the design of a CNN model based on transfer learning concept .
For example, when industrial robots are applied to pick and place tasks of resin molded articles, information of each object’s position and orientation is essential. Recognition and extraction of the object position in an image are not so difficult if some image processing technique is used, however, that of orientation is not easy due to the variety in shape. In this paper, a pick and place robot is introduced while implementing a visual feedback control and a transfer learning-based CNN. The visual feedback control enables to omit the complicated calibration between image and robot coordinate systems, also the transfer learning based CNN allows the robot to estimate the orientation of target objects. The effectiveness and validity of the system is demonstrated through pick and place experiments using a small articulated robot named DOBOT.
 F. Nagata, K. Tokuno, K. Nakashima, A. Otsuka, T. Ikeda, H. Ochi, K. Watanabe, M.K. Habib, Fusion method of convolutional neural network and support vector machine for high accuracy anomaly detection, Procs. of the 2019 IEEE International Conference on Mechatronics and Automation (ICMA 2019), pp. 970-975, Tianjin, China, August 2019.
 F. Nagata, K. Tokuno, K. Mitarai, A. Otsuka, T. Ikeda, H. Ochi, K. Watanabe, M.K. Habib, Defect detection method using deep convolutional neural network, support vector machine and template matching techniques, Artificial Life and Robotics, Vol. 24, No. 4, pp 512-519, 2019.
 F. Nagata, K. Miki, Y. Imahashi, K. Nakashima, K. Tokuno, A. Otsuka, K. Watanabe and M.K. Habib, Orientation detection using a CNN designed by transfer learning of AlexNet, Procs. of the 8th IIAE International Conference on Industrial Application Engineering (ICIAE2020), pp. 295-299, 2020.
Keywords: Pick and place robot, Convolutional neural network, Transfer learning, Visual feedback control
Rohini DeshpandeSchool of Electronics and Communication Engineering, REVA UNIVERSITY, Bengaluru, India
Speech Title: Innovative Design Techniques for Notch Filters
Abstract: In this talk I am proposing two innovative application specific FIR notch filters that are free from most of the undesired features such as non-optimal design, ripple in pass bands, highly involved mathematical computations etc. 1. Maximally flat, linear phase FIR notch filter with controlled null width. First part of the talk deals with design of a maximally flat, linear phase FIR notch filter with controlled null width. Design analyses carried out with first, third and fifth order zero derivative constraints of the amplitude response of the FIR filter at notch frequency is discussed. Detailed analysis in the research work discussed shows that the null width of a maximally flat, linear phase FIR notch filter can be controlled by suitable selection of individual zero odd order derivatives and also by the successive addition of zero odd order derivatives at the notch frequency ωd 2. FIR notch filter with highly narrow rejection bandwidth. Design of FIR notch filters (NF) with highly narrow rejection bandwidth (RBW) is discussed. Reduction in the RBW can be achieved progressively in three stages. In the first stage an FIR notch filter is designed from a second order IIR prototype filter. For a given length L of the NF, the maximum permissible value of ‘r’ (the pole length of IIR prototype filter) is chosen to achieve very narrow RBW of the FIR filter. In the next stage by using an Amplitude Change Function (ACF):H(z)(2 – H(z)), the designed filter is sharpened. Consequently, the RBW of the resulting NF is reduced to almost half of the earlier value. In the third stage, RBW is further reduced by repeated sharpening of the filter by the same ACF.
Tadeusz CzachorskiProfessor, Silesian University of Technology, Poland
Speech Title: Dynamics of Software Defined Networks investigated via diffusion approximation queueing models
Abstract: Network structures based on static switches are not well suited for the needs of high performance, energy efficiency and reliability in dynamically changing environments, and are not flexible enough to maintain Quality of Service (QoS) for increasingly complex networks.
This may be changed by Software Defined Networks which use intelligent programmable controllers being aware of the overall state of nodes and links,
and able to dynamically manage the network and adapt it to new conditions.
They provide flexible and scalable routing by separating the control and data planes for traffic engineering, link failure recovery, load balancing and security issues.
Since standard queueing models are not well adapted to analyse the transient regime, we propose a tractable diffusion approximation for both the transient and steady-state behaviour. The model may represent any network topology transmitting time-dependent flows with routing changes. It computes queue length and delay distributions at each nework node and along complete paths between senders and receivers.
For realistic router parameters, we show that transients states occupy a large fraction of time. Thus network optimisation conducted with SDN controllers should include the effect of time-dependent behaviours. Detailed numerical examples are presented.
Yutaka FukuchiAssociate Professor, Tokyo University of Science, Japan
Speech Title: Rational harmonic mode-locking operation of bismuth-based fiber laser source
Abstract: Wavelength-tunable high-repetition-rate optical pulse generation is essential for many applications such as high-speed optical communication systems and optical signal processing. Harmonically mode-locked fiber lasers have proven to be able to generate wavelength-tunable short pulses with small timing jitter and gigahertz repetition rates. However, maximum repetition rate of the harmonically mode-locked fiber lasers is normally limited by the bandwidth of the intra-cavity modulator and the operating frequency range of the drive electronics. Recently, a scheme called the rational harmonic mode-locking has been proposed and demonstrated to generate optical pulses at very high repetition rates by driving the laser system at a frequency slightly offset from one of its harmonics. In this presentation, we review a stable and wavelength-tunable rational harmonic mode-locked short-cavity fiber laser employing a 1.5-m-long bismuth-oxide-based erbium-doped fiber and a 2.5-m-long bismuth-oxide-based highly nonlinear fiber. Since the short bismuth-oxide-based erbium-doped fiber has a wide gain profile, continuous wavelength tuning covering both the C-band and the L-band can be achieved. The pulse-amplitude equalization can also be achieved by adjusting the bias level and the modulation depth of the intra-cavity modulator. The bismuth-oxide-based highly nonlinear fiber has an ultra-high nonlinear coefficient of 1100 W-1km-1, about 50-60 times higher figure than those of the conventional silica-based highly nonlinear fibers. Because of this feature, the length of the nonlinear fiber for effective suppression of the supermode noise is dramatically shortened; the total cavity length is as short as 10 m. Thus, stable and amplitude-equalized optical pulses up to 40 GHz are successfully obtained for the entire tunable wavelength range.
Teruo KankiAssociate Professor, Osaka University, The institute of Scientific and Industrial Research, Japan
Speech Title: Research on Proton-diffusion in single crystal vanadium dioxides
Abstract: Vanadium dioxide (VO2) is a strongly correlated electronic material with a metal-insulator transition (MIT) over room temperature. Ion-doping to VO2 dramatically alters its transport properties and the MIT temperature. Recently, insulating hydrogenated VO2 (HVO2) accompanied by a crystal structure transformation from VO2 was experimentally observed. Despite the important steps taken towards realizing novel applications, fundamental physics such as the diffusion constant of intercalated protons and the crystal transformation energy between VO2 and HVO2 are still lacking. In this symposium I will talk investigation of the physical parameters of proton diffusion constants accompanied by VO2 to HVO2 crystal transformation with temperature variation and their transformation energies. It was found that protons could propagate several micrometers with a crystal transformation between VO2 and HVO2. The proton diffusion speed from HVO2 to VO2 was approximately two orders higher than that from VO2 to HVO2. The long-range propagation of protons leads to the possibility of realizing novel iontronic applications and energy devices.
Moisés ToapantaProfessor, Department Computer Science, Salesian Polytechnic University (UPS), Ecuador
Speech Title: Analysis for the Adoption of Security Standards to Improve the Management of Securities in Public Organizations
Abstract: Public organizations have the ongoing task of properly managing the security of the information they handle. The objective of this research is to analyze the security standards adopted by public organizations in Ecuador to improve their management of information security. The deductive method was applied for the review and analysis of appropriate standards for public institutions. As a result, information was obtained on the different security policies, standards and guidelines that apply, national and international public organizations. A Diagram of activities for the adoption of standards for public organizations resulted; a prototype standards-based Information Security Management Model; and an Information Security Management Matrix, from which the Risk Mitigation Percentage was calculated. It was concluded that maintaining high levels of security in public organizations requires the adoption of control standards in different areas and the collaboration of the different organizational and hierarchical levels of public organizations.
Tsuyoshi MinamiProfessor, Institute of Industrial Science, The University of Tokyo, Japan
Speech Title: Extended gate-type organic thin-film transistors as chemical sensing platforms for healthcare applications
Abstract: Organic thin-film transistors (OTFTs) have attractive features including lightweight, mechanical flexibility, compact integration. Utilizing such features, we have employed extended gate-type OTFTs functionalized by bio- or artificial materials toward new sensing platforms [1-3]. Due to molecular recognition phenomena, the electrical properties (e.g. drain current, threshold voltage, and transfer or output characteristics) of the OTFT devices vary upon by the addition of the analyte on the extended-gate. Based on the working principle, we have successfully detected organic/inorganic cations, anions, nonions and biomacromolecules . It should be noted that the OTFTs offer simple analytical methods without pre-treatments and complicated procedures. As representatives, we herein demonstrate our successful detection of lactate and immunoglobulins by the OTFTs (Fig. 1A, B) [2,3]. Furthermore, the OTFTs can be integrated with microfluidic systems  for real-time monitoring in biological fluids. We realized the real-time monitoring of changes in glucose (Glc) concentrations (Fig.1C) . Thus, the OTFTs are suitable as sensor devices for the analyses of biological molecules. Overall, our approach would be reliable sensing platforms in practical analyses in the fields of biochemistry and clinical chemistry.
Fig.1 Photograph of the fabricated OTFT. (A) Left: Schematic illustration of OTFT enzyme sensor for lactate. Right: Time-course of drain current with increasing concentration of lactate. (B) Left: Immobilization of anti-IgG antibody on the extended-gate electrode. Right: Transfer characteristics (IDS–VGS) of the OTFT upon titration with IgG in a D-PBS solution with 0.1 wt% bovine serum albumin (BSA). Inset: [IgG] = 0–100 μg/mL. (C) Left: Top view of the OTFT integrated with the microfluidic system for glucose detection. Middle: Side view of the integrated device. Right: Time-course of the drain current with randomly changing concentration of Glc (pseudo glucose consumption by living cells).
Keywords: Organic Thin-Film Transistor, Molecular Recognition, Microfluidic System, Lactate, Immunoglobulin, Glucose
References:  T. Minami et al., ACS Sens., 2019, 4, 2571 (Front Cover);  T. Minami et al., Anal. Sci., 2019, 35, 103;  T. Minami et al., Appl. Phys. Lett., 2014, 104, 243703 (The most accessed article in 2014-2016);  T. Minami et al., ChemElectroChem, 2020, 7, 1332 (Cover Feature)
Noman IslamProfessor, Iqra University, Pakistan
Speech Title: An LSTM based approach for Load Balancing in Cloud Computing
Abstract: The notion of cloud computing is used for providing service as a utility over the internet via dynamic provisioning of resources. Cloud computing suffers from several issues that need to be resolved before its true realization. Load balancing is one of those challenges that requires significant research consideration. In this talk, we will talk about load balancing challenges in cloud computing. The talk presents a taxonomy of those approaches. We will also present a long short term memory (LSTM) based approach for load balancing in cloud environments. Deep learning has been used as a crosscutting solution for several research problems. Using LSTM, one can model long term dependencies between current and past data. This modelling of future state has been utilized and this talk will present a solution that predicts future state of the VM. Based on that, one can proactively migrate VM in case of a possible overload prediction.
Abdullh HasanatAssociate Professor, Computer Engineering, Department of Electrical and Computer Engineering, University of Nizwa, Oman
Speech Title: Probabilistic Home-based Routing Scheme for Delay Tolerant Networks
Abstract: In traditional Mobile Adhoc Networks (MANET) routing algorithms, the existence of a sustainable path between the source and destination is a crucial issue. Despite the rapid advancement in Internet-of-Things (IoT) technologies, these algorithms are shown to be unsuccessful in Delay Tolerant Networks (DTNs) since the network is almost disrupted, and then such a path is not guaranteed. Although Store-and-Forward routing approach has provided an adequate solution for these networks, it suffers from the large end-to-end delay and the additional traffic and communication costs incurring the network. In this talk, I will present the most recent work on routing in DTNs. Then, I will introduce an efficient single-copy home-based routing scheme for DTNs via two different scenarios, blind and semi-blind, based on the amount of available knowledge about the destination. Simulation results presented in this Speech will demonstrate how efficient the presented scheme is with reference to the Epidemic, the unrestricted flooding algorithm. Moreover, compared to state-of-art algorithms, the presented scheme showed tremendous performance improvement in terms of communication cost, end-to-end delay and successful delivery ratio.
Uyen T. NguyenAssociate Professor, Lassonde School of Engineering, York University, Canada.
Speech Title: Online Social Networks: Malware Threats and Countermeasures in Relation to Topological Network Properties
Abstract: The popularity and diverse uses of online social networks (OSNs) have given incentives to hackers and cybercriminals to carry out attacks using malicious software (malware). Given very large populations on social networks (e.g., 2.6 billion users on Facebook), a successful attack can result in tens of millions of user profiles being compromised and computers and devices being infected. This presentation provides an overview of malware threats in OSNs and countermeasures. In this talk, I will first present topological properties of OSNs, namely, low average network distances, power-law distributed node degrees, and high clustering coefficients. I will then discuss different types of malware targeting OSN users as well as countermeasures against OSN malware. The presentation also highlights how the topological properties of OSNs affect the propagation of
malware and design of countermeasures.