NOVEL ADAPTIVE MULTIUSER DETECTIONALGORITHM BASED ON WAVELET TRANSFORM

The wavelet transform-based adaptive multiuser detection algorithm is presented. The novel adaptive multiuser detection algorithm uses the wavelet transform for the preprocessing, and wavelet-transformed signal uses LMS algorithm to implement the adaptive multiuser detection. The algorithm makes use of wavelet transform to divide the wavelet space, which shows that the wavelet transform has a better decorrelation ability and leads to better convergence. White noise can be wiped off under the wavelet transform according to different characteristics of signal and white noise under the wavelet transform. Theoretical analyses and simulations demonstrate that the algorithm converges faster than the conventional adaptive multiuser detection algorithm, and has the better performance. Simulation results reveal that the algorithm convergence relates to the wavelet base, and show that the algorithm convergence gets better with the increasing of regularity for the same series of the wavelet base. Finally the algorithm shows that it can be easily implemented.

A Novel High-Speed Equalizer for QAM Signals

A high-speed equalizer based on a new algorithm： stop-and-go-DD-LMS CMA （SGLMS-CMA） for quadrature amplitude modulation （QAM） signals ispresented. It integrates conventional constant modulus algorithm （CMA） and decision-direct least-mean-square （DD-LMS） under stop-and-go principle. Matlab simulations indicate that, compared with conventional CMA,the new algorithm performs five times faster in convergence speed, 3-5dB improved in rudimental mean square error （MSE）, 82% decreased in operation complexity and can correct a final phase ambiguity. As to the equalizer block in the system,synthesis results show that the SGLMS-CMA ＋ DD-LMS equalizer＇s hardware consumption is only 5% greater than the CMA＋ DD-LMS equalizer＇ s. Finally by using SMIC 0.18μm library to synthesis, the new equalizer is embedded into QAM demodulation chip,and test results show that the new equalizer acts better.

Convergence rate of kernel canonical correlation analysis

Kernel canonical correlation analysis(CCA) is a nonlinear extension of CCA,which aims at extract-ing the information shared by two random variables. It has wide applications in many fields,such as information retrieval. This paper gives the convergence rate analysis of kernel CCA under some approximation conditions and some suggestions on how to choose the regularization parameter. The result shows that the convergence rate only depends on two parameters:the rate of regularization parameter and the decay rate of eigenvalues of compact operator VY X,and it gives better understanding of kernel CCA.

SIGNAL ESTIMATION WITH BINARY-VALUED SENSORS

This paper introduces several algorithms for signal estimation using binary-valued outputsensing.The main idea is derived from the empirical measure approach for quantized identification,which has been shown to be convergent and asymptotically efficient when the unknown parametersare constants.Signal estimation under binary-valued observations must take into consideration oftime varying variables.Typical empirical measure based algorithms are modified with exponentialweighting and threshold adaptation to accommodate time-varying natures of the signals.Without anyinformation on signal generators,the authors establish estimation algorithms,interaction between noisereduction by averaging and signal tracking,convergence rates,and asymptotic efficiency.A thresholdadaptation algorithm is introduced.Its convergence and convergence rates are analyzed by using theODE method for stochastic approximation problems.