DC/DC converter using the proportional-integral (PI) sliding mode control (SMC) scheme is investigated, including the selection of the switching surface, the proof of the reaching condition and the existence condition...DC/DC converter using the proportional-integral (PI) sliding mode control (SMC) scheme is investigated, including the selection of the switching surface, the proof of the reaching condition and the existence condition of sliding motion. The sliding regime and the local stability are given. The implementation of the PI SMC is simpler than other SMC schemes and the steady-state error is eliminated. A prototype based on Buck converter is built up. The experimental results show that the dynamic performance and robustness to the parameter variations and external disturbances are improved.展开更多
Discrete-time chaotic circuit implementations of a tent map and a Bernoulli map using switched-current （SI） techniques are presented. The two circuits can be constructed with 16 MOSFET＇s and 2 capacitors. The simul...Discrete-time chaotic circuit implementations of a tent map and a Bernoulli map using switched-current （SI） techniques are presented. The two circuits can be constructed with 16 MOSFET＇s and 2 capacitors. The simulations and experiments built with commercially available IC＇s for the circuits have demonstrated the validity of the circuit designs. The experiment results also indicate that the proposed circuits are integrable by a standard CMOS technology. The implementations are useful for studies and applications of chaos.展开更多
As a kind of statistical method, the technique of Hidden Markov Model （HMM） is widely used for speech recognition. In order to train the HMM to be more effective with much less amount of data, the Subspace Distribut...As a kind of statistical method, the technique of Hidden Markov Model （HMM） is widely used for speech recognition. In order to train the HMM to be more effective with much less amount of data, the Subspace Distribution Clustering Hidden Markov Model （SDCHMM）, derived from the Continuous Density Hidden Markov Model （CDHMM）, is introduced. With parameter tying, a new method to train SDCHMMs is described. Compared with the conventional training method, an SDCHMM recognizer trained by means of the new method achieves higher accuracy and speed. Experiment results show that the SDCHMM recognizer outperforms the CDHMM recognizer on speech recognition of Chinese digits.展开更多
This paper proposes a method of error detection based on macroblock （MB） types for video transmission. For decoded inter MBs, the absolute values of received residues are accumulated. At the same time, the intra tex...This paper proposes a method of error detection based on macroblock （MB） types for video transmission. For decoded inter MBs, the absolute values of received residues are accumulated. At the same time, the intra textural complexity of the current MB is estimated by that of the motion compensated reference block. We compare the inter residue with the intra textural complexity. If the inter residue is larger than the intra textural complexity by a predefined threshold, the MB is considered to be erroneous and errors are concealed. For decoded intra MBs, the connective smoothness of the current MB with neighboring MBs is tested to find erroneous MBs. Simulation results show that the new method can remove those seriously-corrupted MBs efficiently. Combined with error concealment, the new method improves the recovered quality at the decoder by about 0.5--1 dB.展开更多
基金Supported by the National Natural Science Foundation of China (No.60372004)
文摘DC/DC converter using the proportional-integral (PI) sliding mode control (SMC) scheme is investigated, including the selection of the switching surface, the proof of the reaching condition and the existence condition of sliding motion. The sliding regime and the local stability are given. The implementation of the PI SMC is simpler than other SMC schemes and the steady-state error is eliminated. A prototype based on Buck converter is built up. The experimental results show that the dynamic performance and robustness to the parameter variations and external disturbances are improved.
基金Supported by the National Natural Science Foundation of China (No.60372004) and Natural Science Foundation of Guangdong Province (No.20820)
文摘Discrete-time chaotic circuit implementations of a tent map and a Bernoulli map using switched-current （SI） techniques are presented. The two circuits can be constructed with 16 MOSFET＇s and 2 capacitors. The simulations and experiments built with commercially available IC＇s for the circuits have demonstrated the validity of the circuit designs. The experiment results also indicate that the proposed circuits are integrable by a standard CMOS technology. The implementations are useful for studies and applications of chaos.
基金Supported by the National Natural Science Foundation of China (No.60172048)
文摘As a kind of statistical method, the technique of Hidden Markov Model （HMM） is widely used for speech recognition. In order to train the HMM to be more effective with much less amount of data, the Subspace Distribution Clustering Hidden Markov Model （SDCHMM）, derived from the Continuous Density Hidden Markov Model （CDHMM）, is introduced. With parameter tying, a new method to train SDCHMMs is described. Compared with the conventional training method, an SDCHMM recognizer trained by means of the new method achieves higher accuracy and speed. Experiment results show that the SDCHMM recognizer outperforms the CDHMM recognizer on speech recognition of Chinese digits.
基金the BK21 Foundation of Korea and the National Natural Science Foundation of China (Grant Nos. 60532060, 60672117 and 60607010)
文摘This paper proposes a method of error detection based on macroblock （MB） types for video transmission. For decoded inter MBs, the absolute values of received residues are accumulated. At the same time, the intra textural complexity of the current MB is estimated by that of the motion compensated reference block. We compare the inter residue with the intra textural complexity. If the inter residue is larger than the intra textural complexity by a predefined threshold, the MB is considered to be erroneous and errors are concealed. For decoded intra MBs, the connective smoothness of the current MB with neighboring MBs is tested to find erroneous MBs. Simulation results show that the new method can remove those seriously-corrupted MBs efficiently. Combined with error concealment, the new method improves the recovered quality at the decoder by about 0.5--1 dB.
