• The Journal of the Acoustical Society of Korea
• /
• v.22 no.3
• /
• pp.188-193
• /
• 2003

Regardless of the distribution of the adaptation data in the testing environment, model-based adaptation methods that have so far been reported in various literature incorporates the adaptation data undiscriminatingly in reducing the mismatch between the training and testing environments. When the amount of data is small and the parameter tying is extensive, adaptation based on outlier data can be detrimental to the performance of the recognizer. The distribution of the adaptation data plays a critical role on the adaptation performance. In order to maximally improve the recognition rate in the testing environment using only a small number of adaptation data, supervised weighted training is applied to the structural maximum a posterior (SMAP) algorithm. We evaluate the performance of the proposed weighted SMAP (WSMAP) and SMAP on TIDIGITS corpus. The proposed WSMAP has been found to perform better for a small amount of data. The general idea of incorporating the distribution of the adaptation data is applicable to other adaptation algorithms.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.23 no.6
• /
• pp.623-632
• /
• 2015

This study proposes a control strategy of the common rail pressure with a fuel injection limitation algorithm to reduce particulate matter (PM) emissions under transient states. The proposed control strategy consists of two parts: injection quantity limitation and rail pressure adaptation. The injection limitation algorithm determines the maximum allowable fuel injection quantity to avoid rich combustion under transient states. The fuel injection quantity is limited by predicting the burned gas rate after combustion; however, the reduced injection quantity leads to deterioration of engine torque. The common rail pressure adaptation strategy is designed to compensate for the reduced engine torque. An increase of the rail pressure under transient states contributes to enhancement of the engine torque as well as reduction of PM emissions by promoting atomization of the injected fuel. The proposed control strategy is validated through engine experiments. The rail pressure adaptation reduced the PM emission by 5-10% and enhanced the engine torque up to 2.5%.

• Journal of Communications and Networks
• /
• v.9 no.3
• /
• pp.282-295
• /
• 2007

Due to the nature that high datarate leads to shorter transmission range, the performance enhancement by high datarate 802.11 WLANs may be degraded when applying high datarate to an 802.11 based multihop ad hoc network. In this paper, we evaluate, through extensive simulations, the performance of multihop ad hoc networks at multiple transmission datarates, in terms of the number of hops between source and destination, throughput, end-to-end delay and packet loss. The study is conducted based on both stationary chain topology and mesh topologies with or without node mobility. From numerical results on network performance based on chain topology, we conclude that there is almost no benefit by applying the highest datarate when the chain length is 6 hops or more. With node mobility in mesh topology, the benefit of using high datarate diminishes at even shorter number of hops. To explore the main reasons for this behavior, analyses on multihop end-to-end throughput and network k-connectivity have been conducted later in the paper, and correspondingly an auto-rate adaptation algorithm has been proposed.

• Proceedings of the IEEK Conference
• /
• 2002.06a
• /
• pp.61-64
• /
• 2002

In this paper, various equalizers are considered to improve the performance of Home Phoneline Networking Alliance (HomePNA) 2.0 system under dispersive channel with intersymbol interference. We evaluate and compare the performances of Recursive Least Squares (RLS) and Least Mean Squares (LMS) adaptation algorithms. Computer simulations show that the equalizers utilizing tile RLS algorithm outperforms the LMS algorithm, especially for the system of high symbol rate and complex constellation.

• Journal of Korean Society of Transportation
• /
• v.22 no.4 s.75
• /
• pp.159-173
• /
• 2004

This study utilizes the fuzzy logic and genetic algorithm to improve the existing incident detection models by addressing the problems associated with "crisp" thresholds and model transferability (applicability). The model's major components were designed to be a set of the fuzzy inference engines, and for the self-adaptation capability the genetic algorithm was introduced in optimization(or training) of the fuzzy membership functions. This approach is often called "the hybrid of fuzzy-genetic algorithm" The model performance was tested and found to be compatible with that of the existing well-recognized models in terms of performance measures such as detection rate, false alarm rate, and detection time. This study was not an effort for simple improvement of the model performance, but an experimental attempt to incorporate new characteristics essential for the incident detection model to be universally applicable for various roadway and traffic conditions. The study results prove that the initial objective of the study was satisfied, and suggest a direction that the future research work in this area must follow.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.6
• /
• pp.2144-2159
• /
• 2015

CISCO VNI predicted an average annual growth rate of 66% for mobile video traffic between 2014 and 2019 and accordingly much academic research related to video streaming has been initiated. In video streaming, Adaptive Bitrate (ABR) is a streaming technique in which a source video is stored on a server at variable encoding rates and each streaming user requests the most appropriate video encoding rate considering their channel capacity. However, these days, ABR related studies are only focusing on real-time rate adaptation omitting energy efficiency though it is one of the most important requirement for mobile devices, which may cause dissatisfaction for streaming users. In this paper, we propose an energy efficient prefetching based dynamic adaptive streaming technique by considering the limited characteristics of the batteries used in mobile devices, in order to reduce the energy waste and provide a similar level of service in terms of the average video rate compared to the latest ABR streaming technique which does not consider the energy consumption. The simulation results is showing that our proposed scheme saves 65~68% of energy at the average global mobile download speed compared to the latest high performance ABR algorithm while providing similar rate adaptation performance.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.4 s.42
• /
• pp.17-25
• /
• 2006

An intelligent game has been studied for some time, and the main purpose of the study was to win against human by enhancing game skills. But some commercial games rather focused on adaptation of the user's behavior in order to bring interests on the games. In this study, we are suggesting an adaptive reinforcement learning algorithm, which focuses on the adaptation of user behavior. We have designed and developed the Othello game, which provides large state spaces. The evaluation of the experiment was done by playing two reinforcement learning algorithms against Min-Max algorithm individually. And the results show that our approach is playing more improved learning rate, than the previous reinforcement learning algorithm.

• The KIPS Transactions:PartB
• /
• v.11B no.2
• /
• pp.199-206
• /
• 2004

This paper proposes a hybrid independent component analysis(ICA) of fixed-point(FP) algorithm and robust algorithm. The FP algorithm is applied for improving the analysis speed and performance, and the robust algorithm is applied for preventing performance degradations by means of very small kurtosis and temporal correlations between components. And the adaptive adaptation of temporal correlations has been proposed for solving limits of the conventional robust algorithm dependent on the maximum time delay. The proposed ICA has been applied to the problems for separating the 4-mixed signals of 500 samples and 10-mixed images of $512\times512$pixels, respectively. The experimental results show that the proposed ICA has a characteristics of adaptively adapting the maximum time delay, and has a superior separation performances(speed, rate) to conventional FP-ICA and hybrid ICA of heuristic correlation. Especially, the proposed ICA gives the larger degree of improvement as the problem size increases.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.1
• /
• pp.47-50
• /
• 2019

This paper proposes an acoustic model adaptation method for effective speech recognition in noisy environments. In the proposed algorithm, the noise corruption function is estimated employing DNN (Deep Neural Network), and the function is applied to the model parameter estimation. The experimental results using the Aurora 2.0 framework and database demonstrate that the proposed model adaptation method shows more effective in known and unknown noisy environments compared to the conventional methods. In particular, the experiments of the unknown environments show 15.87 % of relative improvement in the average of WER (Word Error Rate).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.6B
• /
• pp.559-568
• /
• 2002

In this paper, we studied on combining the control of transmission fates and power for the real system where a finite set of transmission rates are used. In [1], the combined control of transmission rates and power was first researched, and suggested the Selective Power Control (SPC) algorithm. However, it can't guarantee the minimum rate to each user and results in frequent changes of rate due to oscillation of the SIR (Signal to Interference Ratio) values. As a main purpose of this paper, we derive a formulation model and propose a distributed iteration algorithm to solve these problems. To evaluate the performance of the proposed algorithm, we carried out numerical analysis and computational experiments. The results indicate that the proposed algorithm achieves better throughput than conventional one by keeping the low average transmission power.