Compared Features of Ultra Wideband with features of existing techniques such as Wi-Fi.Analyzed UWB transmission system (transmitter, receiver, and transmission channel), UWB positioning system based on UWB transmission system, and UWB-based Wireless Positioning Estimated Algorithms..Simulated positioning model in 2-dimension and 3-dimension systems, and improved average positioning error distance to 7cm

Used u-law algorithm to compress original speech data from 16-bit to 8-bit data, and Used IMA ADPCM encoder to encode compressed data.Used IMA ADPCM decoder to decode compressed data, and used u-law algorithm to expand and to reconstruct new speech data. New speech data can clearly express the same content

Conducted individual study on the Echo Cancellation and Adaptive Algorithms (PNLMS, Mu-law, and NLMS). Learned LMS, PNLMS, NLMS Algorithms and Echo Cancellation. Achieved NLMS, PNLMS, MPNLMS, and SPNLMS learning curves by MATLAB. Compared the advantages and disadvantages of four algorithms for small and large coefficients.

Used ASM method to detect and to crop a specic face from a raw photo with multiple people. Then, used Gabor filter to the cropped face in order to do edge detection. Next, used PCA method to extract its face features. Trained a machine learning model using data stored in the database (10 pictures per people), and compared face features to do face recognition. Corrected rate to detect a specic face from a raw multi-people photo is 89%

Modeded the texture by a convolution of a 2D excitation process with a 2D FIR. Found three parameters: Elongation, Orientation, ML estimator and compare them with cost function to find best way to synthesis new image. Successfully repeated Dr. Eom’s “2D Moving Average Models for Texture Synthesis and Analysis”

Compared multiple signal borders extension methods in repairing damaged image. Achieved image edge detection with wavelets.