• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.5
• /
• pp.489-496
• /
• 2006

The random-based cleaning algorithm is a simple algorithm widely used in commercial vacuum cleaning robots. This algorithm has two limitations, that is, cleaning takes a long time and there is no guarantee that the cleaning will cover the whole cleaning area. This has lead to customer dissatisfaction. Thus, in recent years, many intelligent cleaning algorithms that takes into consideration information gathered from the cleaning area environment have been proposed. The plowing-based algorithm, which is the most efficient algorithm known to date when there are no obstacles in the cleaning area, has a deficiency that when obstacle prevail, its performance is not guaranteed. In this paper, we propose the Group-k algorithm that is efficient for that situation, that is, when obstacle prevail. The goal is not to complete the cleaning as soon as possible, but to clean the majority of the cleaning area as fast as possible. The motivation behind this is that areas close to obstacles are usually difficult for robots to handle, and hence, many require human assistance anyway In our approach, obstacles are grouped by the complexity of the obstacles, which we refer to as 'complex rank', and then decide the cleaning route based on this complex rank. Results from our simulation-based experiments show that although the cleaning completion time takes longer than the plowing-based algorithm, the Group-k algorithm cleans the majority of the cleaning area faster than the plowing algorithm.

• Smart Structures and Systems
• /
• v.16 no.6
• /
• pp.1023-1047
• /
• 2015

In this study, the optimal location of the MR fluid segments in a partially treated laminated composite sandwich plate has been identified to maximize the natural frequencies and the loss factors. The finite element formulation is used to derive the governing differential equations of motion for a partially treated laminated composite sandwich plate embedded with MR fluid and rubber material as the core layer and laminated composite plate as the face layers. An optimization problem is formulated and solved by combining finite element analysis (FEA) and genetic algorithm (GA) to obtain the optimal locations to yield maximum natural frequency and loss factor corresponding to first five modes of flexural vibration of the sandwich plate with various combinations of weighting factors under various boundary conditions. The proposed methodology is validated by comparing the natural frequencies evaluated at optimal locations of MR fluid pockets identified through GA coupled with FEA and the experimental measurements. The converged results suggest that the optimal location of MR fluid pockets is strongly influenced not only by the boundary conditions and modes of vibrations but also by the objectives of maximization of natural frequency and loss factors either individually or combined. The optimal layout could be useful to apply the MR fluid pockets at critical components of large structure to realize more efficient and compact vibration control mechanism with variable damping.

• Journal of the Korea Society for Simulation
• /
• v.14 no.2
• /
• pp.93-103
• /
• 2005

This paper presents characteristics of a virtual human's navigation using a limited perception-based mapping. Previous approaches to virtual human navigation have used an omniscient perception requiring full layout of a virtual environment in advance. However, these approaches have a limitation on being a fundamental solution for a human-likeness of a virtual human, because behaviors of humans are basically based on their limited perception instead of omniscient perception. In this paper, we integrated Hill's mapping algorithm with a virtual human to experiment virtual human's navigation with the limited perception. This approach does not require full layout of the virtual environment, 360-degree's field of view, and vision through walls. In addition to static objects such as buildings, we consider enemy emergence that can affect an virtual human's navigation. The enemy emergence is used as the variable on the experiment of this present research. As the number of enemies varies, the changes in arrival rate and time taken to reach the goal position were observed. The virtual human navigates by two conditions. One is to take the shortest path to the goal position, and the other is to avoid enemies when the virtual human encounters them. The acquired result indicates that the virtual human's navigation corresponds to a human cognitive process, and thus this research can be a framework for human-likeness of virtual humans.

• Journal of the Korean Institute of Educational Facilities
• /
• v.23 no.3
• /
• pp.3-10
• /
• 2016

This study compared evacuation effectiveness between the conventional school floor plan and a suggested hypothetical floor plan that was generated by a mathematical model that is commonly applied in the field of industrial engineering. Recent school buildings became much more complicated in floor planning due to the new educational system and modern curriculum than the old days. Nevertheless, architect's approach to floor composition in terms of fire emergency evacuation planning, still has no relation to optimized effective but relies more on a conventional school planning. Therefore, since elementary school buildings are much more likely to be exposed to any fire related events than middle school or high-school, emergency exit effectiveness based on spatial composition has to be seriously evaluated with a scientific method. The algorithm, which acquires the number of persons in each spatial type (node) per floor and the minimum physical distance between spatial types (arc), can propose the most optimized spatial compositing per floor regarding emergency evacuation event. Consequently, this study evaluated elementary school's fire exit effectiveness with the scientific tool and suggested the most reliable spatial composition per floor regarding possibile emergency evacuation event.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.18 no.11
• /
• pp.24-34
• /
• 2019

An automotive automatic transmission is a popular power-transmitting device in passenger vehicles, as it provides various speed ratios for diverse driving conditions with easy manipulation and smooth gear shifting. The transmission is mainly composed of input and output shafts, planetary gear sets, brakes/clutches, and housing, and it yields multiple forward gears and one reverse gear by actuating the shifting devices of the brakes and clutches. In developing a new transmission, kinematic configurations of a transmission, which presents a brief structure and actuation schemes for speed ratios, need to be checked to determine if the structure can be assembled in a layout. It is impossible for a transmission concept having any interference in connecting main components to be developed further in the design process, since connection interference leads to failure of a layout design in the 2-D plane. In this research, an analysis of the assemblability of a front-wheel drive automatic transmission is carried out on an example concept design by applying the vertex addition algorithm based on graph theory.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.39 no.7
• /
• pp.89-96
• /
• 2002

In this paper, the ADI(Adaptive De-interlacing) algorithm is proposed, which improves visually and subjectively horizontal and vertical edges of the image processed by the ELA(Edge Line-based Average) method. This paper also proposes a VLSI architecture for the proposed algorithm and the architecture designed through the full custom CMOS layout process. The proposed algorithm is verified using C and Matlab and implemented using $0.6{\mu}m$ 2-poly 3-metal CMOS standard libraries. For the circuit and logic simulation, Cadence tool is used.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.13 no.2
• /
• pp.140-146
• /
• 2013

In this paper, we describe a vision sensor-based driving algorithm for indoor automatic guided vehicles (AGVs) that facilitates a path tracking task using two mono cameras for navigation. One camera is mounted on vehicle to observe the environment and to detect markers in front of the vehicle. The other camera is attached so the view is perpendicular to the floor, which compensates for the distance between the wheels and markers. The angle and distance from the center of the two wheels to the center of marker are also obtained using these two cameras. We propose five movement patterns for AGVs to guarantee smooth performance during path tracking: starting, moving straight, pre-turning, left/right turning, and stopping. This driving algorithm based on two vision sensors gives greater flexibility to AGVs, including easy layout change, autonomy, and even economy. The algorithm was validated in an experiment using a two-wheeled mobile robot.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.11
• /
• pp.2902-2912
• /
• 1994

The ultrasonic array sensor system, consisting of one transmitter and fourreceivers instead of the traditional combination of a transmitter and a receiver is proposed in order to identify the location of objects. From the theoretical analysis and the experimental results, it is found that this new array sensor system could derive the information on the position of objects accurately, while the traditional sensor system could provide only the informatioin on the distance to objects. This sensor system is used to develop a sonar-based local mapping algorithm. The local map is used to find the existence of possible gates, through which the mobile robots can pass, and to select the suitable one in order for the robots to reach the goal safely in the presence of obstacles. The performance of the proposed local map algorithm is demonstrated experimentally in a small working area with several obstacles. It is found that the quality of the resulting local map is sufficient for the avoidance of collisions between the robots and obstacles and for the selection of the suitable gate leading to the goal. It is also shown that the global map of the working area could be obtained by integrating several local maps constructed from different locations and that it matches the actual layout of the working area well.

• Smart Structures and Systems
• /
• v.15 no.2
• /
• pp.355-373
• /
• 2015

An optimal sensor placement (OSP) method based on structural subsection technique (SST) and model reduction technique was proposed for modal identification of truss structures, which was conducted using genetic algorithm (GA). The constraints of GA variables were determined by SST in advance. Subsequently, according to model reduction technique, the optimal group of master degrees of freedom and the optimal objective function value were obtained using GA in a case of the given number of sensors. Correspondingly, the optimal number of sensors was determined according to optimal objective function values in cases of the different number of sensors. The proposed method was applied on a scaled jacket offshore platform to get its optimal number of sensors and the corresponding optimal sensor layout. Then modal kinetic energy and modal assurance criterion were adopted to evaluate vibration energy and mode independence property. The experiment was also conducted to verify the effectiveness of the selected optimal sensor layout. The results showed that experimental modes agreed reasonably well with numerical results. Moreover the influence of the proposed method using different optimal algorithms and model reduction technique on optimal results was also compared. The results showed that the influence was very little.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.48 no.1
• /
• pp.31-38
• /
• 2011

With the increasing demand for high-speed transmission systems, adaptive equalizers have been widely used in receivers to overcome the limited bandwidth of channels. In order to reduce the cost for testing high-speed receiver chips, on-chip eye-opening monitoring (EOM) technique which measures the eye-opening of data waveform inside the chip can be employed. In this paper, a 10-Gb/s adaptive 2-tap look-ahead decision feedback equalizer (DFE) with EOM function is proposed. The proposed EOM circuit can be applied to look-ahead DFEs while existing EOM techniques cannot. The magnitudes of the post-cursors are measured by monitoring the eye of received signal, and coefficients of DFE are calculated using them by proposed adaptation algorithm. The circuit designed in 90nm CMOS technology and the algorithm are verified with post-layout simulation. The DFE core occupies $110{\times}95{\mu}m^2$ and consumes 11mW in 1.2V supply voltage.