This article discusses automatic strategy acquisition for the game "Othello" based on a reinforcement learning scheme. In our approach, a computer player, which initially knows only the game rules, becomes stronger after playing several thousands of games against another player. In each game, the computer player refines the evaluation function for the game state, which is achieved by min-max reinforcement learning (MMRL). MMRL is a simple learning scheme that uses the min-max strategy. Since the state space of Othello is huge, we employ a normalized Gaussian network (NGnet) to represent the evaluation function. As a result, the computer player becomes strong enough to beat a player employing a heuristic strategy. This article experimentally shows that MMRL is better than TD(0) and also shows that the NGnet is better than a multi-layered perceptron, in our Othello task.

This paper introduces an acoustic echo canceller system materialized with a 16-bit fixed point processing type DSP (Analog Devices, ADSP-2181). This experimental system uses the tri-quantized-x individually normalized least mean square (INLMS) algorithm little degrading the convergence property under the fixed point processing. The experimental system also applies a small step gain to the algorithm to prevent the double-talk from increasing the estimation error. Such a small step gain naturally reduces the convergence speed. The experimental system compensates the reduction by applying the block length adjustment technique to the algorithm. This technique enables to ceaselessly update the coefficients of the adaptive filter even when the reference signal power is low. The experimental system thus keeps the echo return loss enhancement (ERLE) high against the double-talk.

A standard of Advanced Storage Magneto Optical (AS-MO) having a 6 Gbyte capacity in a 120 mm-diameter single side disk was established by using a magnetically induced superresolution readout method. Transition from in-plane to perpendicular magnetization for exchange-coupled readout layer (GdFeCo) and in-plane magnetization mask layer (GdFe) of the AS-MO disk has been investigated using the noncontinuous model. The readout resolution was sensitive to the thickness of the readout layer. To evaluate readout characteristics of AS-MO disks, the simulation using micro magnetics model was performed and the readout layers were designed. The readout characteristics of the AS-MO disk is improved by making the readout layer thinner.

The effect of surface roughness is crucial for contact recording and proximity recording. In this paper a probability model is developed for investigation of the influence of surface roughness on flying performance and the contact force of the slider. Simulations are conducted for both the contact recording slider and the proximity recording slider, and the results are well coordinated with the reported experimental results and the self-conducted experimental results. Studies are further extended to the characterization of the roughness of the air bearing surface and the disk surface that may support head/disk spacing between 5 nm and 15 nm.

Both frequency- and time-domain analyses of glide signals from a PZT glide-slider flying over a laser zone-textured (LZT) thin film disk medium were used to determine the slider vibration at a small disk-slider clearance. Slider vibration was found to be particularly dependent on the uniformly placed laser bump and the effects due to the air-bearing stiffness over the LZT medium. We found that a high density of small, pointed laser bumps (10X) has a more distinct impact on airflow than large, jagged-rim craterlike laser bumps (1X) on the slider. We therefore investigated the effect of laser bump density on the slider vibration, and found that marginally higher laser bump density (3X versus 2X) results in higher slider vibration. While resonant vibration has been a major glide problem, the effects of laser bump density have also recently become important in the face of ultralow glide height, 0.5 µ" (12 nm). Its influence can be clearly observed when the disk-slider clearance becomes very small. At such an ultrasmall disk-slider clearance, even minimal slider vibration can be detrimental to the head-disk interface. Taking into account the various contributions of slider vibration and considering possible damage to the head-disk interface, it is clear that the optimization of laser bump design should go beyond just the glide height and coefficient of stiction. It should take into account the effects of laser bump height, density and spatial distribution on vibration-induced flying height variation while maintaining a low glide height and coefficient of stiction. An ideal LZT medium should therefore have low bump height to enable low glide height, i. e. , 0.5 µ" (12 nm), but specific bump shapes and sufficient density to achieve low stiction. Laser bump density should, however, be controlled to moderate its effect on slider vibration and possibly disk-slider collision (297 words).

Thermal stability of anisotropic and isotropic Co alloy thin-film media is investigated. The orientation ratio of CoCrTa(Pt)/Cr media was controlled by the mechanical texture of the NiP/Al substrates. Bulk magnetic properties, delta M curves and time decay of magnetization in the circumferential and radial directions were measured. The maximum magnetic viscosity coefficient calculated from the time decay of magnetization in the circumferential direction was higher than that in the radial direction for a mechanically textured sample, while it was similar in both directions for a non-textured sample. The magnetic viscosity coefficient in the circumferential direction is smaller than that in the radial direction when the reverse field is in the range of the demagnetization field for thin-film recording media. This implies that an anisotropic sample (namely, a sample with a high orientation ratio) will be more thermally stable when it is not exposed to a large external magnetic field.

This paper describes the design and implementation of a multi-threaded Distributed Shared Memory (DSM) system, called Cohesion, which provides high programming flexibility and latency masking, and supports load balancing. Cohesion offers a parallel programming environment which is very similar to that on a multiprocessors system. Threads could be created recursively in this environment, and users are not required to handle the locations of the threads. Instead of supporting a shared variable model, Cohesion provides a global shared address space among all nodes in the system. The space is further divided into three regions, i. e. , release, conventional, and object-based memory, each is applied with different consistency protocol. In this paper, the design issues in an ordinary thread system, such as thread management, load balancing, and synchronization, have been reconsidered with the memory management provided by the DSM system. Several real applications have been used to evaluate the performance of the system. The results show that multi-threading usually has better performance than single-threading because the network latency can be masked by overlapping communication and computation. However, the gain depends on program behavior and the number of threads executed on each node in the system.

This paper proposes an adaptive list-output Viterbi equalizer (LVE) with fast compare-select operation, in order to achieve a good trade-off between bit error rate (BER) performance and processing speed. An LVE, which keeps several survivors for each state, has good BER performance in the presence of wide-spread intersymbol interference. However, the LVE suffers from large processing delay due to its sorting-based compare-select operation. The proposed adaptive LVE greatly reduces its processing delay, because it simplifies compare-select operation. In addition, computer simulation shows that the proposed LVE causes only slight BER performance degradation due to its simplification of compare-select operation. Thus, the proposed LVE achieves better BER performance than decision-feedback sequence estimation (DFSE) without an increase in processing delay.

Image coding with vector quantization (VQ) reveals several defects which include edge degradation and high encoding complexity. This paper presents an edge-preserving coding system based on VQ to overcome these defects. A signal processing unit first classifies image blocks into low-activity or high-activity class. A high-activity block is then decomposed into a smoothing factor, a bit-plane and a smoother (lower variance) block. These outputs can be more efficiently encoded by VQ with lower distortion. A set of visual patterns is used to encode the bit-planes by binary vector quantization. We also develop a modified search-order coding to further reduce the redundancy of quantization indexes. Simulation results show that the proposed algorithm achieves much better perceptual quality with higher compression ratio and significant lower computational complexity, as compared to the direct VQ.

This paper proposes a new polling-based dynamic slot assignment (DSA) scheme. With the rapid progress of wireless access systems, wireless data communication will become more and more attractive. In wireless data communication, an efficient DSA scheme is required to enhance system throughput, since the capacity of radio links is often smaller than that of wired links. A polling-based DSA scheme is typically used in centralized slot assignment control systems. It, however, is difficult to assign the slots to the targeted mobile terminals in a fair-share manner if only a polling-based scheme is used, especially in unbalanced-traffic circumstances, as revealed later. To solve this problem, we propose the exponential decreasing and proportional increasing rate control as is employed in available bit rate (ABR) service in ATM so that fair slot assignment is achieved even in heavily-unbalanced-traffic circumstances. Moreover, so that an AP operating with a large number of MTs can avoid long transmission delays, a polling-based resource request scheme with random access is featured in a new algorithm. Simulations verify that the proposed scheme offers fair slot assignment for each user while maintaining high throughput and short delay performance.

This paper proposes a novel spread spectrum (SS) modem for random access satellite communication systems that employs digital matched filters. The proposed modem employs a parallel structure to ensure detection of packet arrival. Code timing detection with a combination of a coarse detector and a fractional error detector reduces the sampling rate while maintaining the BER performance. An in-symbol pilot multiplexing scheme is also proposed for fast and stable carrier synchronization with a simple hardware. A performance evaluation shows that the proposed modem achieves the UW miss-detection probability of 10-4 at the Eb/No of 0 dB. The overall BER performance achieved in experiments well agrees simulation.

The constant modulus algorithm (CMA) of the adaptive array has been developed for suppressing the co-channel interference and the intersymbol interference in mobile communications. In this paper a novel CMA for the hybrid of the adaptive array and equalizer (HAE) is proposed to combat the problems of insufficient degrees of freedom and mainbeam multipath interferers. The HAE with CMA utilizes the constant modulus property for the output signal of the HAE to adjust the weight vectors of the array and equalizer simultaneously. The co-channel interferers can be canceled by the array and the multipath interferers can be removed by the array or the equalizer following the array in the HAE. Therefore, the array in the HAE with CMA may need less number of elements than that required by the CMA array which cancels both the co-channel interferers and multipath interferers. Besides, the presence of the mainbeam multipath interferers, which may seriously degrade the performance of the CMA array, has much less effect on the HAE with CMA since it can be suppressed by the equalizer instead of the array. Simulation results are presented to demonstrate the merits of the CMA for the HAE.

A semi-synchronous circuit is a circuit in which every register is ticked by a clock periodically, but not necessarily simultaneously. In a semi-synchronous circuit, the minimum delay between registers may be critical with respect to the clock period of the circuit, while it does not affect the clock period of an ordinary synchronous circuit. In this paper, we discuss a delay insertion method which makes such a semi-synchronous circuit faster. The maximum delay-to-register ratio over the cycles in the circuit gives a lower bound of the clock period. We show that this bound is achieved in the semi-synchronous framework by the proposing gate-level delay insertion method.

This paper presents a novel implication-based method for logic minimization in large-scale, multi-level networks. It significantly reduces network size through repeated addition and removal of redundant subnetworks, utilizing multi-signal implications and relationships among these implications. These are handled on a transitive implication graph, proposed in this paper, which offers the practical use of implications for logic minimization. The proposed method holds great promise for the achievement of an interactive logic design environment for large-scale networks.

This paper proposes an ASIP performance optimization method taking clock frequency into account. The performance of an instruction set processor can be measured using the execution time of an application program, which can be determined by the clock cycles to perform the application program divided by the applied clock frequency. Therefore, the clock frequency should also be tuned in order to maximize the performance of the processor under the given design constraints. Experimental results show that the proposed method determines an optimal combination of FUs considering clock frequency.

In several design methods for Pass-transistor Logic (PTL) circuits, Boolean functions are expressed as OBDDs in decomposed form and then the component OBDDs are directly mapped to PTL cells. The total size of OBDDs (number of nodes) corresponds to the circuit size. In this paper, we investigate a method for PTL synthesis based on exact minimization of Free BDDs (FBDDs). FBDDs are well-studied extension of OBDDs with free variable ordering on each path. We present statistics showing that more than 56% of 616126 NPN-equivalence classes of 5-variable Boolean functions have minimum FBDDs with less size than their OBDDs. This result can be used for PTL synthesis as libraries. We also applied the exact minimization algorithm of FBDDs to the minimization of subcircuits in the synthesis for MCNC benchmarks and found up to 5% size reduction.

This letter is concerned with a new algorithm which can be used to design a time-domain equalizer (TEQ) for xDSL systems employing the discrete multitone (DMT) modulation. The proposed algorithm, derived by neglecting the terms which do not affect the performance of a DMT system in ARMA modeling, is shown to have a good performance compared with the previous TEQ algorithms even with a significantly lower computational complexity. In addition, the proposed algorithm does not require the channel impulse response or training sequence, since all processing is made only with the received data.

This paper describes a novel layout optimization technique using electromagnetic (EM) simulation. Simple equivalent circuits fitted to EM simulation results are employed in this method, to present a modification guide for a layout pattern. Fitting errors are also investigated with some layout patterns in order to clarify the applicable range of the method, because the errors restrict the range. The method has been successfully adopted to an X-band low noise MMIC amplifier (LNA). The layout pattern of the amplifier was optimized in only two days and the amplifier has achieved target performances--a 35 dB gain and a 1.7 dB noise figure--in one development cycle. The effective chip area has been miniaturized to 4.8 mm2. The area could be smaller than 70% in comparison with a conventional layout MMIC.

We propose a new technique that is able to extract the small-signal equivalent-circuit elements of high electron mobility transistors (HEMTs) without causing any gate degradation. For the determination of extrinsic resistance values, unlike other conventional techniques, the proposed technique does not require an additional relationship for the resistances. For the extraction of extrinsic inductance values, the technique uses the R-estimate, which is known to be more robust relative to the measurement errors than the commonly used least-squares regression. Additionally, we suggest an improved cold HEMT model that seems to be more general than conventional cold HEMT models. With the use of the improved cold HEMT model, the proposed technique extracts the extrinsic resistance and inductance values.

This paper describes a new method for detecting the gaze position of a user on a monitor from monocular images. In order to detect the gaze position, we extract facial features (both eyes, nostrils and lip corners) automatically in 2D camera images and estimate the 3D depth information and the initial 3D positions of those features by recursive estimation algorithm in starting images. Then, when a user moves his/her head in order to gaze at one position on a monitor, the moved 3D positions of those features can be estimated from 3D motion estimation by Extended Kalman Filter (EKF) and affine transform. Finally, the gaze position on a monitor is calculated from the normal vector of the plane determined by those moved 3D positions of features. Especially, in order to obtain the exact 3D depth and positions of initial feature points, we unify three coordinate systems (face, monitor and camera coordinate system) based on perspective transformation. As experimental results, the 3D depth and the position estimation error of initial feature points, which is the RMS error between the estimated initial 3D feature positions and the real positions (measured by 3D position tracker sensor) is about 1.28 cm (0.75 cm in X axis, 0.85 cm in Y axis, 0.6 cm in Z axis) and the 3D motion estimation errors of feature points by Extended Kalman Filter (EKF) are about 3.6 degrees and 1.4 cm in rotation and translation, respectively. From that, we can obtain the gaze position on a monitor (17 inches) and the gaze position accuracy between the calculated positions and the real ones is about 2.1 inches of RMS error.