Keynote Speakers

Witold PedryczName:Witold Pedrycz
Affiliation:University of Alberta, Canada
Title:Multiview Rule-based Modeling and Granular Aggregation

Abstract:  Multiview models, as the name stipulates, are models describing a real-world system observed from different points of view. In establishing such individual perspectives, we typically engage locally available features (attributes, input variables). When considered together, a collection of multiview models has to be aggregated. Multiview models also arise in the presence of data coming with a massive number of variables when building a monolithic model involving all attributes is neither feasible nor computationally sound.
We formulate and discuss these two categories of scenarios by focusing on fuzzy rule-based architectures. An important task when building an aggregate of multiview models is to endow the overall global model with a sound measure of quality using which one can efficiently assess the relevance of the individual results produced by the rule-based models as well as establish the quality of the overall fusion. We advocate that the quality of the results can be quantified in terms of some information granule. In the two scenarios outlined above, the family of multiview models is aggregated with the use of the augmented principle of justifiable granularity -one of the fundamentals of Granular Computing. The related optimization criteria of coverage and specificity are discussed along with the associated optimization process. The emergence of type-2 information granules is also shown.

Makoto IwasakiName:

Makoto Iwasaki 

Affiliation:Nagoya Institute of Technology, Japan

Modeling and Identification of Strain Wave Gearing for Motion Control Applications to Precision Positioning Devices

Abstract:  The invited speech presents a practical motion controller design technique for precision positioning devices including strain wave gearing, e.g. industrial multi-axis robots. Since HarmonicDrive® gears (HDGs), a typical strain wave gearing, inherently possess nonlinear properties, such as Angular Transmission Errors (ATEs), nonlinear stiffness, nonlinear friction etc., due to structural errors and flexibility in the mechanisms, the ideal positioning accuracy corresponding to the apparent resolution cannot be essentially attained at the output of gearing in the devices. In addition, mechanisms with HDGs generally excite resonant vibrations due to the periodical disturbance by ATEs, especially in the condition that the frequency of synchronous components of ATE corresponds to the critical mechanical resonant frequency. This speech, therefore, focuses on the accurate settling and vibration suppression in positioning, in order to improve the performance deteriorations by applying several control approaches, e.g. a model-based feedforward control manner, and adaptive/robust feedback control manners. In the speech, at first, an approach toward modeling and identification of the ATEs is presented to achieve and apply the precise mathematical models to the motion controller design. Following to the accurate mathematical modeling, several motion controller design approaches are practically presented: a model-based feedforward compensation to improve the settling performance in the positioning, an adaptive variable notch filter design as a feedback compensation to effectively suppress the mechanical vibration, and an H_inf controller design to boost the robust capability. The proposed approaches have been applied to precision motion control of actual devices as servo actuators, and verified through numerical simulations and experiments

Jinseok WOOName:

Jinseok WOO 

Affiliation:Tokyo University of Technology, Japan

Development Direction of Human Coexistence Robot Partner based on Smart Device

Abstract:  Social isolation can cause problems on human both mentally and physically. In particular, the isolation of the elderly causes serious problems such as lonely death. However, social participation can lead to healthier lives by reducing isolation. Many technologies are being developed to reduce social isolation and loneliness. Accordingly, we considered to use robot partner in order to prevent social isolation. In this paper, we explain the current state about our development of robot partner, and we discuss development direction of robot. First, we show our robot partner system. Next, we explain the modular structured systems in order to develop robot partner system. Finally, we show several examples of the robot system, and discuss the applicability of proposed system.

Xianlin ZengName:

Xianlin Zeng 

Affiliation:Beijing Institute of Technology, China

Accelerated First-Order Distributed Method for Nash Equilibria of Convex-Concave Bilinear Two-Network Zero-sum Games

Abstract:  Due to the big data/scale of many modern applications and rise of distributed optimization/game, simple first-order gradient descent algorithms have become dominant and have been particularly revisited because gradients are often computed in a cheap and distributed way. Many researchers focus on the development of accelerated first-order algorithms. However, most existing works focus on unconstrained optimization problems. Convex-concave bilinear two-network zero-sum games have applications in several domains such as machine learning, optimization, and game theory. This talk will introduce a recent result on a dynamical accelerated first-order method for Nash equilibria seeking of a class of two-network zero-sum games. The proposed method achieves an O(1/t2) convergence rate in some sense, which is consistent to Nesterov’s accelerated method for unconstrained optimization problems.

Xin ChenName:

Xin Chen 

Affiliation:China University of Geosciences, China

Technology and application of intelligent humanoid robot system

Abstract:  After recent advancement of computing and robotics technologies, intelligent robot system spread widely all walks of life. However, these robots face several typical engineering challenges. First of all, the intelligent perception of the environment is the basis of robot operation. Secondly, the robot has real-time decision-making ability to external information. Finally, high precision control can be performed to achieve the target operation. Two typical robot systems, live working robot system and dulcimer intelligent playing robot system, are introduced in detail in this report. A complex workspace environment perception technology put forward by combining broad-spectrum light source and binocular vision to realize high-precision positioning of different sizes of operation targets. A humanoid behavior learning model is proposed, which can realize intelligent real-time decision-making of dual-arm performance by learning the behavior habits of human professional players. A humanoid trajectory planning method is introduced to realize the anthropomorphic motion of two arms with joint constraint. Some preliminary results are demonstrated via video.