Search

Quantitative structure-activity relationship (QSAR) regression models are mathematical ones which relate the structural properties of chemicals to the potencies of the biological activities of the chemicals. In QSAR models, the physical and chemical information of the molecules is encoded into quantitative numbers called descriptors. Recently, experimental test results (profiles) have been used as descriptors of chemicals. Profile QSAR 2.0 (pQSAR) model suggested by Martin et al. , is a multitask, two step machine learning prediction method with a combination of random forest regressions (RFRs) and partial least squares regression (PLSR). In pQSAR model, one fills the profile table’s missing values with RFRs and then builds PLSR using the profile predictions. Note that in the second step of the pQSAR method, PLSR’s predictor variables are profiles; so activity values, and the response variables are also activity values. Thus we can use the PLSRs to update the profile table and then repeat the second step. In this work, we propose an extended model of pQSAR generated by RFRs and PLSRs. Experiment of updating the given full initially predicted profile table by two kinds of prediction models, RFRs and PLSRs, has been conducted iteratively for the PKIS and ChEMBL data sets. Even though prediction performance of individual combination of RFRs and PLSRs varies, the average of the all possible predicted profile tables for given iteration shows better performance. This ensemble model has better prediction performance in sense of Pearson’s $R^{2}$ compared to that of the pQSAR model.

Recent studies on quantum computing algorithms focus on excavating features of quantum computers which have potential for contributing to computational model enhancements. Among various approaches, quantum annealing methods effectively parallelize quadratic unconstrained binary optimization (QUBO) formulations of systems of linear equations. In this paper, we simplify these formulations by exploiting congruence of real symmetric matrices to diagonal matrices. We further exhibit computational merits of the proposed QUBO models, which can outperform classical algorithms such as QR and SVD decomposition.

A Non-dispersive Infrared (NDIR) sensor for carbon monoxide (CO) was used to investigate the effects of physical and chemical factors on the accuracy of the sensor. The sensor's measurement range was 0.05 – 200 ppm. It was found that the measurement error of the NDIR at 50 ppm of CO was the lowest. Moreover, the increase of concentrations of the interferential gases (CO2, NO, NO2) led to the decrease of the NDIR accuracies. It was observed that the temperature of the inlet gas did not significantly affect NDIR measurement. On the contrary, the accuracy of the devices was gradually declined with the increase of humidity at 50 ppm of CO.

In this study, the formation of a toroidal microvortex by optoelectrokinetic effect was numerically simulated using COMSOL v4.2a multiphysics software. AC voltage was applied to the two parallel electrodes in a microchannel to generate temperature gradient in the fluids. In addition to the AC electrothermal effect, local heating by a laser illumination was also considered. Numerical simulations were conducted for dielectric fluids. The toroidal microvortex induced by the optoelectrokinetic effect shows that two counter-rotating vortices are produced above the bottom electrodes. Fluid motions in the middle of bottom boundary are cancelled out by flows in opposite directions and consequently producing stagnation. It is expected that micro/nano particles are deposited in bottom electrode. Local heating enhanced the intensity of microvortex substantially due to the additional temperature gradient, it was confirmed that the AC electrothermal effect with laser illumination can be used for rapid concentration of micro/nano particles in the spot area.

Gas atomization is the conventional and well-known methods to produce metal powders. This method can be used to get high sphericity and fluidity of the powders. The metal powders produced by the gas atomization have a rapid cooling rate because of fast heat transfer between the droplets and the injected gases. The gas atomization can be used to produce various metal powders with melting temperatures below 1700°C such as aluminum and copper alloys, nickel alloys, and stainless steels etc. Usually the stainless steel powders for metal injection mould process have been also produced by the gas atomization. In this paper, we produced the SUS 439 using the gas atomization process. Due to the high strength and good corrosion resistance of the SUS 439, it has been used widely in various applications, especially in automobile, military and aerospace instruments. As-prepared gas atomized powders were observed through visual identification method. Microstructure of the powers was observed by optical microscopy, size and size distribution and shape were observed by FE-SEM and LPSA. Phase of the powers was analyzed by XRD. Sintering behaviors were checked also with sintering conditions.

The metal porous body has interesting structural characteristics such as high specific strength and lightness and high antivibration, damping and sound absorption properties and also has excellent functional characteristics such as unique thermal insulation and catalytic carrier absorption and so on. With these advantages, recently, it has been using a high performance lubricant bearing, filters for filtration, absorbing materials, and heat exchanger and fuel cell components materials etc. Therefore, it has been actively researched to produce the metal porous body having the good properties. There are some several methods to make the porous metal body. Among them, the PU dipping process is excellent, simple and economical effective method. In this study, the porous Fe foam by the PU dipping process was prepared and characterized. Fe slurry was prepared by dispersing Fe powders in the distilled water (DIW), in where the dispersing agent hydroxyalkyl methyl cellulose (Hecellose) was used. After then the porous polyurethane template was dipped in the Fe slurry followed with drying in the air for 1 day under room temperature. The porous metal body was obtained after burning out the porous polyurethane at 600°C in the argon atmosphere for 1 hour followed with sintering at 1200°C in the argon atmosphere for 1 hour. In the first step of sintering process, pyrolyzing and debinding was happened to remove out the polyurethane. The prepared porous metal's appearance was observed through visual identification method. Microstructure of Fe foam was observed by optical microscopy, pore size and shape were observed by FE-SEM, and so on.

Polyaniline-copper (PANI-Cu) and polyaniline-silver (PANI-Ag) composites were prepared by electrical explosion of wire (EEW) in presence and the absence of polyacrylic acid (PAA) surfactant. Structure characteristics of the composites were studied by field emission scanning electron microscope (FE-SEM), X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), UV-vis spectroscopy and electrical measurement. The conductivity of polyaniline-metal increased several magnitude compared to pure PANI. The characteristics of composites changed significantly when PAA was used. The stability of composite colloids increased; the surface of composite was smoother and the conductivity decreased. PANI-Ag composite have higher conductivity than PANI-Cu.

In this paper, the auxiliary switch control using lookup table is proposed to enhance the efficiency of a bi-directional DC/DC converter for EVs. The bi-directional DC/DC converter performs soft switching. It is difficult to obtain continuous current data at the resonant operation. The proposed auxiliary switch control has a look-up table reference of the auxiliary switch turn-on time. The control method properly controls the auxiliary switch turn-on time according to the required load current and is sufficiently safe and efficient in a regenerative mode. The validity of the proposed approach has been demonstrated by the experimental results on a bi-directional DC/DC converter, inverter, IPMSM.

In this work, polyaniline - copper (PANI-Cu) and polyaniline/polyacrylic acid - copper (PANI/PAA-Cu) composites were synthesized by electrical explosion of wire (EEW) in the liquid. The PANI and PAA polymers in ethanol were stirred for 24 hrs. and sonicated for 2 hrs. The polymer solutions used as a liquid for explosion process to fabricate the Cu nanoparticle. X-ray diffraction (XRD) results of the composites showed that the Cu was polycrystal. Field emission scanning electron microscope (FE-SEM) was used to observe the microstructure of the composites. The PANI particles were in round shape and covered on the as-prepared Cu nanoparticles. The Cu nanoparticles were spherical and dispersed well in the polymer matrix. Crystalline size of the Cu nanoparticles was about 25nm. The obtained composites were extensively characterized by Fourier Transform Infrared (FTIR) spectroscopy, and electrical measurements.

Flow rate control is the uppermost concern for trochoid hydraulic pump. Cavitation within the flow field of pump has the most influence on flow rate control reason at approximately 3500 ∼ 4000 RPM high speed rotation of pump. In this paper, based on AMESim® and CATIA®, we will present how to simulate cavitation by analyzing the control factors of trochoid pump; hydraulic pressure change of outlet, flow rate according to rotation speed of inner rotor, leakage through gap between outer rotor and inner rotor, and discharging angle of outlet. The proposed methodology of cavitation simulation will enables field engineers to have access to the design of trochoid pump more easily and thereby to have more concrete control over the flow rate of pump by realizing its analysis model similar to its actual product model.

Wind energy is one of the most promising renewable energy resources. BAWTs have been studied recently. BAWTs are a new technology trend to combine wind turbines with buildings and structures and it has several advantages compared with the conventional wind turbine such as minimization of electrical transmission loss and construction cost due to the distributed power sources and using buildings. In this paper, 110 kW class horizontal axis wind turbine is designed and CFD analysis is carried out with various wind speed. The results show aerodynamic power output of 110 kW class BAWT is higher than that of 110 kW class stand alone wind turbine due to the concentration effect caused by the wind speed acceleration between buildings. Flow characteristics are quite different for the existence of 110 kW class BAWT. These results can be applied to the research and development of BAWTs and high performance computing is indispensable.

It is our interest to design and develop a high power pulser system employing microplasma discharge in a spark gap system. The type of discharge generates short electric pulses capable of both the characteristics of high repetition rate and fast rise time. Such a switch pulser has wide applications in the industry. An efficient electrical discharge in the spark gap is very important for the switching operation. The regime of operation utilizes the residual plasma in the inter pulse recovery of the spark gap to increase the repetition rate. Therefore, the mechanisms of the discharge in a controlled manner, such as the plasma discharge by the arc channel, the dielectric recovery process as well as residual plasma in the post arc period are of great importance to ensure the high repetition rate. The aforementioned mechanisms are established through optimization of the circuit element parameters such as the electrode geometry, the gap distance, the gas pressure, the gas type as well as the applied voltage and current ratings.

To control the quality of air pollutants, several key air pollutants are measured and monitored offline and also online by tele-monitoring (TMS) system which is built for the management of air quality. Until now, most of the monitoring methods has been used in an univariate approach which measures and monitors a single pollutant, such as particular matter (PM2.5 or PM10). Since it has a correlation relationship between variables, this study focuses on a multivariate statistical monitoring method for monitoring the indoor air quality in a subway. The proposed method consists of three main components: (1) principal component analysis (PCA) to reduce the dimensionality of multivariate air pollutant data and to remove collinearity; (2) global model to monitor nine real-time air pollution data and diagnose the status of indoor air quality during one year; (3) local models to keep track of the seasonal variations of air pollutants. The seasonal models are suggested to consider the variations of the pollutant concentration according to the climate change in Korea. The multivariate monitoring method is applied to a real time TMS dataset of nine air pollutants in a real subway station. It shows the accurate and reliable result of air pollutants in a subway over univariate monitoring, which can significantly enhance the power of the monitoring system. And the seasonal models allow to isolate the characteristics of the seasonal variations for specific monitoring of air pollutants. The multivariate approach is useful to check the indoor air quality status using all information of the sensors and to predict effects of indoor air pollutants to the passengerpsilas health.

Micromixer has been studied for several years. In this paper we have attempted to design and develop a new type of combinatorial chaotic and serpentine micromixer (CSC) that is compatible with the microfabrication process of polydimethylsiloxane (PDMS). We visualize the mixing of the fabricated CSC micromixer. A mixing experiment was performed using the Rhodamine B with Micro-LIF. Such a new and novel micro-mixer is simple to fabricate and easy to use. The micromixer is characterized and visualized with the help of the Micro-LIF technique. The new micromixer will e further applied to lab-on-chip device. The mixing capability of this mixer is about 30-33%.

This paper presents the fabrication of nano-porous polymer microcapsules using self-assembly of the block copolymers in the microchannel. The spherical microcapsule with the diameter of about 25 mum was successfully fabricated with the thickness distribution of 0.5-1.5 mum and nano pores. Its potential as an encapsulant use was demonstrated by comparing the release performance under the various microenvironments.

We present Sensos, a sensor node operating system with the device management scheme for sensor nodes with multiple sensor elements. As the number of sensor elements attached to a single sensor node increases, programming of applications for each sensor elements becomes a significant burden to the developer. A device management scheme providing well-defined APIs independent to a specific sensor element itself can make the program development more efficient and convenient. The basic concept of the device management scheme integrated in this proposed sensor node operating system, Sensos, is based on IEEE 1451 standard. We redefined and simplified the existing standard to make it applicable to the sensor node operating system. Implementation of Sensos shows a remarkable small footprint less than 55 Kbytes and we demonstrate that applications for new sensor devices can be easily developed using the proposed Sensos

This paper presents the centrifugation using the self-rotating secondary flow in the microchamber. The feasibility of the microfluidic centrifuge was demonstrated numerically an experimentally. The centrifugation of the nano-particles was visualized and its performance was characterized quantitatively. Also, a new finding of the transient motion of the rotating flow influenced the stable operation of the centrifuge.

Reconfiguration of WDM networks allows network operators to reconfigure the topologies in response to changing traffic demands to improve the QoS. To reconfigure the current virtual topology, the time to reconfigure is decided, and then a target topology is created. Next, real transition between topologies is occurred. We propose a VTR controller for WDM networks. First, our controller decides when current virtual topology is reconfigured. Some traffic may travel with single hop lightpaths and others may not. What the amount of traffic which is serviced through multi-hop increases shows that current virtual topology is not adequate for traffic demands. We use this information to reconfigure. And, the controller creates a target topology. As discontinuous wavelengths increases in networks, it can increase the difficulty of establishing single hop lightpaths for future traffic. Hence, the controller establishes the lightpaths which would remain discontinuous as few as possible after assigning.

In this paper, we present a new VHDL intermediate format CDFG (Control Data Flow Graph) and a minimal hardware resource scheduling algorithm for ASIC design automation. The intermediate format, CDFG represents the constraints which limit hardware design such as conditional branch, sequential operation and time constraints. The proposed scheduling algorithm could handle the conditional branches effectively and could check the timing constraints efficiently. It minimizes the total operating time by reducing the number of the constraints as maximal as possible, determining the number of control steps in minimum bound. The effectiveness of the proposed algorithm has been proven by the experiment with the benchmark examples.

We report on the observation of LaGuerre-Gaussian (LG) beam profiles of photonic quantum ring (PQR) hole-type lasers. Naturally formed LG beams are highly valuable for optical tweezing applications, like transporting or manipulating DNA molecules. Various devices with different injection currents and hole diameters have been analyzed, some of which show promising results.

Single-walled carbon nanotube biosensor was developed for the detection of Escherichia coli using aptamer-functionalized carbon nanotube field effect transistors. RNA-based E. coli aptamer was chosen by the SELEX process, and immobilized on sidewalls of carbon nanotubes by CDI-Tween 20 linking molecules. The electrical response of the carbon nanotube sensor before and after the binding of E. coli aptamer and E. coli were studied.

In this study, the Ni-based (Ni59Zr20Ti16Si2Sn3) BMG powders were produced by the high pressure gas atomization process, and the ductile Cu powders were mixed using the spray drying process or the ball milling process. The Ni-based amorphous powders and Cu added composite powders were compacted by a simple die hot press and a spark plasma sintering (SPS) processes into cylindrical. In the hot press, the prepared powders were pressed at 570degC (in the at Tg-Tx range) to avoid the crystallization. However, the compacted samples did not reach the theoretical density. At the 900degC, we could get the full densified specimens, but the crystallization of the samples happened. The relative density increased from 79% to 98% when the pressure increased up to 350 MPa by optimum control of the SPS process parameters. The micro-Vickers hardness was also increased from 774 Hv to 810 Hv with increasing temperature from 450degC to 460degC.

Metallic (silver and tungsten) nanopowders were produced by the high current electrical explosion of wire (EEW) under Ar atmosphere. Most of particles have a size of

In the present work, the microstructure and properties of Cu-based composites with 4.5 vol.% dispersed TiB2 particles developed by a combined process of mechanical alloying (MA), self-propagate high temperature synthesis (SHS) and spark-plasma sintering (SPS) will be reported. The microstructure of the composites revealed that the TiB2 dispersoids with the particle size of smaller than 250 nm uniformly distributed in Cu matrix. The sintered samples obtained approximately 99% in relative density after sintering at 650 degC for 5 min. The fine dispersed TiB2 dispersoids in Cu matrix was the main reason to improve the mechanical properties of the composites. The hardness reached a value of 74 HRB. The result of tensile test showed that the ultimate tensile strength, yield strength and Young modulus increased from 187 MPa, 158 Mpa and 119 GPa for pure Cu to 485.7 MPa, 355.3 MPa and 157 GPa for the composites, whereas the electrical conductivity remained a high value of 83 % IACS. After soaking at 850, 900 and 950 degC, the hardness of composites decreased to 62, 61 and 55 HRB, respectively. The softening temperature was determined to be higher than 900 degC.

A simple maximum likelihood detector is described for system-on-chip (SoC) implementation for optical disk drives. As compared to the Viterbi detector (VD) commonly used for storage devices, this detector offers a solution for the speed bottleneck of the VD and less hardware complexity. Using look-ahead samples eliminates the need for a path memory and also resolves the speed bottleneck so that only a "select" operation is required within a clock cycle by moving the other two operations "add" and "compare" out of a clock cycle loop. Another advantage of this detector is that the target sample values can be constructed at the current sample time. This means that sample error can be estimated with significantly less delay for adaptive control loops. This detector is based on two-dimensional boundary functions, which can be readily derived by a simple procedure for finding a set of error events in the respective signal transition and envelope groups. This detector is currently used in a Samsung DVD super-multi-SoC

We have designed and fabricated a microneedle array with electrical functionality with the final goal of electroporating skin's epidermal cells to increase their transfection by DNA vaccines. The microneedle array was made of polymethylmethacrylate (PMMA) by micromolding technology from a master PDMS mold, followed by metal deposition, patterning using laser ablation, and electrodeposition. This microneedle array possessed sufficient mechanical strength to penetrate human skin in vivo and was also able to electroporate red blood cells as an in vitro model to demonstrate cell membrane permeabilization.

This paper outlines the work performed to date in our unit on precision bone fragment positioning during craniofacial surgical procedures. The effort began with the development of computer-aided precision planning of such procedures. Optimization of landmark positions on mobile fragments was performed to best fit a normal form in first two-dimensional then threedimensional models. Work is in progress to extend such optimizations to 3D CT based surface models. Execution of these precision plans was first performed using intraoperative caliper measurement of landmark movements. Intermediate occlusal splints are our current standard positioning device. Work in progress involves automated vision systems and a passive robot for such purposes.

GaN-based laser diodes for 405 nm high-power applications are demonstrated. By decreasing the Al concentration of n-cladding layers, the vertical divergence angle was reduced to 300 mW.

A reconfigurable mesh consists of a mesh-connected topology augmented by a dynamically reconfigurable bus system. Efficient utilization of processing resources in a large, multiuser parallel system depends on the reliable processor management scheme. We present a dynamic and reliable processor allocation strategy to increase the performance of mesh-connected parallel systems with faulty processors. The idea is to reconfigure a faulty mesh system into a maximum convex system using the fault-free boundary nodes to compensate for the nonboundary faulty nodes. To utilize the nonrectangular shaped system parts, our strategy tries to allocate L-shaped submeshes instead of signaling the allocation failure. Extensive simulations show that the strategy performs more efficiently than other strategies in terms of the completion time, the job response time and the system utilization.

This study establishes a numerical model of a three junction GTO device. A three-junction configuration provides good visualization of the model's performance and understanding of the significance of the parameters. Modeling methods are provided using the micromodel method and macromodel method. The problems to be considered for the snubber design such as voltage spike reduction, maximum GTO anode current, and switching power are discussed using the calculation model. As result of the calculations, the snubber circuit stray inductance is found to play an important role in optimizing the minimum switching power loss. Thus, the proposed model GTO device model in this paper solves conventionally difficult problems which cannot have simulation results since convergence cannot be accomplished well in the simulation process and hence cannot treat a simulation completely. >

The study proposes a fuzzy analytic algorithm for classification of mixed pixel (mixel) image. First, the misclassified problem and information loss of mixed pixels are examined. To solve the mixel's uncertainty, a fuzzy function which produces a fuzzy number is suggested. Second, algorithms for the analytic fuzzy approach are represented. This analytic method is applied to two bridges across the Hon River and an area adjacent to the river in Seoul to verify it. >

We present a multicriteria fuzzy controller (MFC) based on fuzzy measures and fuzzy integrals. The basic idea underlying this approach is based on analyzing the source of attributes of the system output responses and applying the fuzzy measure and integral theory to the existing fuzzy controllers. With this scheme, we can tune the three attributes, including rise time, overshoot, and settling time of the output responses. We demonstrate that MFCs have superior control performance compared to conventional fuzzy controllers by computer simulations as well as via experiments performed on a DC servomotor position controller. Moreover, our scheme can be easily implemented in practice simply by adding weight terms to an existing fuzzy rule system and assigning the weight to each rule by employing the fuzzy measure and integral. >

We present a detailed analysis of the performance of queueing schemes in the Batcher-banyan networks with multiple outlets. Each queueing scheme is analyzed under different traffic distributions. In the Batcher-banyan networks with higher switching speedup, the performance of buffering techniques affect the switching system performance and are also related to input and output traffic distributions. Therefore, we investigate queueing schemes for several Batcher-banyan networks using our proposed analysis model. Numerical results indicate that the queueing performance such as switching throughput, latency, and packet loss rate for several Batcher-banyan networks are closely affected by input traffic distribution.

In this paper, we describe a design methodology for fuzzy inference systems based on regression trees. Fuzzy associative memory banks with symmetric triangular fuzzy membership functions are constructed by using sample data statistics of regression tree partitions. We propose a new fuzzy inference method that takes the membership of the average of normalized input features in the antecedent of a production rule.

The aim of this study is to propose an efficient method for partition of spectral space into fuzzy subspace for multi-spectral remote sensing image. The suggested method predicates on sequential subdivision of the fuzzy subspace, and the size of constructed fuzzy space is variable. Under this procedure, n-dimensional pattern space, after considering the distributional characteristic patterns, is partitioned into two different fuzzy subspaces. From the two fuzzy subspaces, the pattern space for further subdivision is chosen; then, this subdivision procedure recursively repeats itself until the stopping condition is fulfilled. The result of this study is applied to 2, 4, 7 band of satellite Landsat TM and satisfactory result is acquired. >

Due to the performance limitations of electrical level circuit simulators, the number of design circuits that cannot be analyzed but need to be simulated at the electrical level are increasing. Hence, extensive research has been carried out with the aim of improving simulation performance for large and complex circuits. This paper introduces a new environment for electrical level simulation with high performance. PESE uses circuit partitioning based on the channel connectedness, in conjunction with a standard electrical simulator like SPICE, to provide a greater simulation speed with less memory requirement while maintaining the desired level of electrical accuracy. The size of circuits PESE can handle is not limited, and the speedup is expected to grow with the circuit size while having much less memory space requirement. Several benchmark circuits have been tested and the results show that PESE provides accurate simulation capability with a speed and memory space advantage.

The cache memory system for CalmRISC32 embedded processor is described in this paper. A dual data cache system structure called a cooperative cache that takes advantage of design flexibilities of a dual cache structure is used as the cache memory system for CalmRISC32 to improve performance and reduce power consumption. The cooperative cache system is applied to both data cache and instruction cache. This paper describes the structure and operational model of the cache memory system for CalmRISC32. The implementation of the cache memory system for CalmRISC32 is also presented.

Architecting today's embedded processor core faces several important design challenges: low power, high performance, and system-on-a-chip considerations. Moreover, support for high-level language constructs and operating systems becomes increasingly critical for acceptance to various applications. CalmRISC/sup TM/-32 effectively meets these challenges by incorporating a carefully designed instruction set, an energy-efficient pipeline design, debugging support with trace mode/CalmBreaker/sup TM/ (an in-circuit debugger), and a generic, yet efficient coprocessor interface. Using a 0.25 /spl mu/m static CMOS standard cell library and compiled datapath cells, the first implementation of CalmRISC/sup TM/-32 operates at 130 MHz (under worst conditions) and consumes 150 /spl mu/A/MHz at 2.5 V. This paper presents a brief description of the instruction set, the overall microarchitecture, and the coprocessor interface of CalmRISC/sup TM/-32.

Processors require both intensive and extensive functional verification in their design phase to satisfy their general purposability. The proposed random vector verification method for CalmRISC/sup TM/-32 core meets this goal by contributing complementary assistance for conventional verification methods. It adopts a cycle-accurate instruction level simulator as a reference model, runs simulation in both the reference and the target HDL and reports errors if any difference is found between them. These processes are automatically performed in the unified environment. The instruction level simulator, the core part in the verification environment is able to simulate almost every aspect of RISC processors from functional behavior of each opcode to timing details in the pipeline flow in fast speed. Its design style from microprogramming scheme also makes its structure modular and flexible.

A low-power cache structure for embedded processors, called a cooperative cache system, is presented in this paper. The cooperative cache system reduces power consumption of the cache system by virtue of the structural characteristics that consists of two separate caches having different associativities and block sizes. The cooperative cache system is adopted as the cache structure for the CalmRISC-32 embedded processor, the prototype chip of which was recently manufactured with a 0.25 /spl mu/m, 4-metal process by Samsung Electronics Co.