Search

Digital watermarking technology is an important technology for copyright protection. In practical applications, the robustness of digital watermarking technology is more demanding. In order to resist attacks and improve invisibility, a new data watermarking scheme is proposed. First, watermark local feature regions will be determined by an improved method which contains two steps: texture complexity roughly location and SIFT precisely location. Second, DCT transformation on these regions is applicated and a set of low frequency coefficients of chosen regions are selected to construct a DCT coefficient matrix. Last, SVD decomposition is performed on DCT coefficient matrix and a QR code carrying watermark information is embedded by the additive rule. The watermark detection algorithm is the reverse process of embedding. The experimental results show that the algorithm has good robustness to common attacks.

Orthogonal frequency division multiplexing (OFDM) is one of the promising techniques in the next generation high-speed data rate wireless networks for its ability to combat the intersymbol interference(ISI), and fair scheduling algorithms are used to utilize the independence of the fading statistics of different user to improve the spectral efficiency. In this paper, we try to analysed the performance of fair scheduling algorithms in the OFDM systems and choose the appropriate subband size to tradeoff between the throughput and the channel feedback. We show that asymptotic scheduling gain of fair scheduling algorithms in the OFDM system increases as the number of user increases and decreases with the average received SNR. We also present the distribution of the averaged received signal-to-noise-ratio (SNR) in a subband, give a formula to evaluate the scheduling gain approximately, and point out that the maximal eigenvalue is a good indicator to choose the appropriate subband size. It is shown that the appropriate subband number is 16 for channel A or 32 for channel B.

A new timing recovery method named prefiltered delayed XOR is presented in this paper, and it is verified that this method is very effective for application in bandlimited multilevel QAM systems compared with other methods. In the mean time, the computer simulation results of the timing jitter feature under various conditions are given, such as the signal to noise ratio, the carrier error and the nonideality of the prefilter. The simulation results are very close to the actual and show that this method is applicable for bandlimited multilevel QAM systems.

An optimum combining algorithm applied in space diversity diffuse channel is presented. Theoretically, this algorithm gives the space diversity channel's performance threshold and achieves the best performance compared to others, such as the equal weight combining algorithm and the selectable combining algorithm. Since the channel bandwidth in a general diffuse channel is relatively narrow, we can only consider the relationship between flat fading and the bit error rate, ignoring the influence of frequency selection fading.

A practical analysis for the quantization noise spectrum of a linear digital filter output estimated from the quantization mean square error is introduced. We give a general quantization mean square error function that considers both roundoff and overflow errors. For a linear filter/system with stationary Gaussian input sequence, the quantization mean square error can be calculated. Then quantization noise level is estimated from the quantization mean square error. The analysis is verified by the design example of a pulse shape filter in CDMA systems.