Design method for a new control system for an autonomous underwater vehicle using linear matrix inequalities

The independent administrative corporation Japan Agency for Marine–Earth Science and Technology (JAMSTEC) has developed a small light autonomous underwater vehicle (AUV) named marine robot experimental 1 (MR-X1).¹ The motion control of MR-X1 is considered in this article. Since the dynamics of MR-X1 mainly depends on its own speed, the motion control is a nonlinear control system. We propose a new controller design method for this system using linear matrix inequalities (LMIs). This algorithm gives a solution as a linear matrix inequality, and can be adapted to solve many LMIs simultaneously. LMIs can be obtained by substituting several speeds into the dynamics of the MR-X1. The proposed controller, which can be derived from the solution of the LMIs, was adapted to MR-X1 and showed good performance in experiments. This work was presented in part at the 11th International Symposium on Artificial Life and Robotics, Oita, Japan, January 23–25, 2006     

Electrostatic probe measurement in an industrial furnace for high-temperature air conditions

Electrostatic probe measurements are reported that identify flame location, displacement speeds of reaction region, and other flame properties within an industrial furnace that is operated with high-temperature preheated air. The electrostatic probe has advantages over other methods when a furnace is operated with high-temperature air. The probe consisted of a fine detection wire and a supporting tube that played a role of the reference electrode. The reaction regions were found to be widely dispersed and weakened as they moved downstream. However, the ion-current signals still included many sharp peaks, perhaps associated with the thin reaction thickness, contrary to the flame structure expected from the high-temperature air combustion. It was also possible to estimate the displacement speeds of reaction region by using the cross-correlation method between two ion current records detected by parallel detection components. The results demonstrate that the electrostatic probe is useful to detect the structure and state of the reaction mode in industrial furnaces even in the presence of high-temperature air combustion.     

Real-time animation of spark discharge

We have developed a CG technique that creates real-time animations of spark discharges. The technique proposed in this paper has factors related to the generation of a discharge, such as electric charge distribution and boundary conditions in a virtual space as the input, and generates spark discharge shape patterns based on the electric field defined by the input information. An electric field is expressed by the Laplace equation. Our method efficiently obtains the numerical solution of the equation using the calculation technique of the conjugate gradient method implemented on GPU, and can cope with dynamic changes in input. It also produces discharge patterns in both two and three dimensions. In addition, an efficient pseudo-dimensional expansion technique is proposed in this paper, which uses multiple two-dimensional electric fields to generate three-dimensional discharge patterns.     

Circuit level prediction of device performance degradation due to negative bias temperature stress

A circuit level methodology for predicting performance degradations due to negative bias temperature stress is developed in this paper. Degradation mechanism is discussed based on experimental observations. Then, models that consist of a threshold voltage shift and a drain current reduction are developed based on the degradation mechanism. The developed models are implemented into a compact MOSFET model so that we can directly link the local degradation of pMOSFETs’ electrical characteristics to the total circuit performances. The validity of the developed models is confirmed by the good agreement in simulated and measured results of I–V characteristics of pMOSFET in all the transistor working region before and after negative bias temperature stress. Then, circuit performance prediction is carried out for the stressed 199-stage ring oscillator on its waveform and oscillation frequency. Excellent agreements between the experimental results and predicted results are obtained.     

Selective pyrolysis of polystyrene to that with a desired low polymeric degree

To obtain low polymeric polystyrene (PS), pyrolysis of high polymeric PS in solution was studied in the temperature range from 290 to 400°C by using additives or acid catalysts. The low polymeric PS targeted here was that with average molecular weight of 10⁴. When the feed PS was pyrolyzed in tetralin by adding sulfur or diphenyl disulfide, the molecular weight of PS decreased greatly, even at lower temperatures, and the desired low polymeric PS was formed in a relatively large amount at the temperatures below 350°C. The degradation behavior was able to be explained in terms of a random polymer chain scission mechanism initiated by sulfur radicals formed from the additives. © 1998 John Wiley & Sons, Inc. J Appl Polym Sci 70: 2299–2305, 1998     

Experimental demonstration and analysis of high performance and low 1/f noise Tri-gate MOSFETs by optimizing device structure

In this paper, we experimentally investigate the performance of multi-gate MOSFETs (MUGFETs) using the advanced radical gate oxide and the accumulation-mode (AM) FD-SOI MOSFETs. Firstly, we experimentally demonstrate that the drain current in AM multi-gate MOSFET is improved about 1.3 times compared with conventional inversion-mode (IM) MOSFETs with the same gate oxide. Secondly, we indicate that 1/f noise levels in AM MUGFETs are obviously suppressed compared with the conventional IM MUGFETs. The advantages resulted from the AM device structure for MUGFETs are demonstrated in this experiment.     

Improvement of critical current density in Y---Ba---Cu---F---O compounds by addition of platinum

Platinum added Y---Ba---Cu---F---O compounds with nominal composition YBa₂Cu₃F_0.4O_xPt_y {y=0, 0.3, 0.5} were prepared by a partial melt process performed at lower temperature than the conventional melt process. Optical micrograph observations showed that Y₂BaCuO₅ (211) particles increase in density and are dispersed finely when increasing the doping of Pt up to 0.3 wt%. For the 0.5 wt% Pt sample, however, the size of the 211 particles became larger than in the samples containing less than 0.3 wt% Pt. The magnetization was measured using a SQUID magnetometer in the temperature range 20–77 K under magnetic fields B_e up to 1 T. The magnetization M also increased with increasing quantity of platinum up to 0.3 wt%. The J_c values estimated from the width of the magnetization curves were 9.8×10⁷ A/m² for the 0.3 wt% Pt sample and 5.4×107 A/m² for the undoped sample at 77 K and B_e=0.8 T. In addition, the whereabouts of the platinum in the quenched samples was investigated by using electron spectroscopy for chemical analysis (ESCA).     

Control of pore size distribution of silica gel through sol-gel process using water soluble polymers as additives

To control the pore size distribution of silica gel, several kinds of water soluble polymers were added in the sol-gel process starting from tetraethyl orthosilicate. The polymers employed here were classified into three groups. The first group comprised uncharged polymers such as polyvinyl alcohol and polyethylene glycol, which affected only the amount of micropores and small mesopores (<20 nm). The second group comprised charged polymers (polyelectrolyte) such as polyacrylic acid (anionic) and a quaternary salt of polyethylene imine (cationic), which significantly increased the number of larger mesopores (>20nm). The third group comprised proteins such as lipase and albumin and decreased the number of micropore and mesopores and greatly increased the number of macropores. The results could almost be explained in terms of the principal effects of the polymers on a process of sol particle growth.     

Separation of linear and non-linear imaging components in high-resolution transmission electron microscope images

Linear and non-linear image components in high-resolution transmission electron microscope images were successfully separated by applying a bandpass filter to the three-dimensional Fourier spectrum of its through-focus series of images. In the observed lattice image of a wedgeshaped Si [110] crystal, we determined the magnitude of the contribution of the non-linear imaging components to the total image intensity distribution. The contribution was proved to become sometimes larger than that of the linear imaging component, even at a thickness of 13 nm.