J.S. Hammerschmidt

Bio: J.S. Hammerschmidt is an academic researcher from Agere Systems. The author has contributed to research in topics: Brute-force search & Metric (mathematics). The author has an hindex of 1, co-authored 1 publications receiving 5 citations.

A maximum a posteriori MIMO detector using recursive metric computations

Abstract: A complexity reduction technique for soft-output maximum a posteriori probability detection in wireless multiple-input multiple-output (MIMO) channels using rectangular quadrature amplitude modulation (QAM) constellations is derived. In contrast to existing low-complexity detection schemes that reduce the number of candidate tests, we instead focus on lowering the computational burden per candidate. The search over a large number of hypotheses to generate output L-values is facilitated by a doubly recursive mechanism, exploiting differential updates of the distance metrics from candidate to candidate. Moreover, the search over all transmit candidates is simplified using a hierarchical "triangular" (zig-zag) shaped candidate scanning order. The technique is derived conceptually using mathematical terms, and then illustrated by means of a 2 times 2 MIMO system using 64-QAM modulation. Under certain practical assumptions, the new scheme is shown to reduce the overall number of operations per candidate by a factor of about 50-100. While the concept is demonstrated by means of the exhaustive search case, it can be combined with other search-based MIMO detection techniques such as spherical decoding

A New Image Denoising Algorithm via Bivariate Shrinkage Based on Quaternion Wavelet Transform

Abstract: Denoising of images corrupted by additive white Gaussian noise is a classical problem in image processing. This paper proposes an efficient algorithm for removing Gaussian noise from corrupted image by using quaternion wavelet transform and bivariate shrinkage function. The image is decomposed by quaternion wavelet transform to obtain the quaternion coefficients in each sub-band. Then a bivariate shrinkage function filter by using the maximum a posteriori is used to model the statistical dependencies among intra-scale quaternion coefficients. The proposed algorithm is compared with other denoising techniques in terms of PSNR. The experimental results indicate that the proposed algorithm outperforms the other denoising algorithms significantly.

Partial MAP-Based List Detection for MIMO Systems

Abstract: The partial maximum a posteriori probability (PMAP) principle can be applied to reduce the complexity of the multiple-input-multiple-output (MIMO) detection through the successive interference cancellation (SIC). In this paper, we apply the PMAP principle to the list detection method for MIMO detection, where the SIC is performed with a list of candidates. The PMAP principle helps to choose candidate symbol vectors in the list detection. It is shown that the proposed method outperforms the conventional list detection method with a reasonable complexity.

A low-complexity max-log-MAP detector

Abstract: A low-complexity soft-output Max-log-MAP (maximum a posteriori probability) detector for N2-QAM is exemplified for a 64-QAM 2timesM multiple-input-multiple-output (MIMO) system, exhibiting the following novel features: (1) hierarchical formulation of the exhaustive metric calculation to reduce the number of candidate tests by a factor of N per received symbol, and (2) multiplier-free implementation of the exhaustive search with 8-fold parallelization. The computational complexity is reduced by a factor of about 250, resulting in a chip area of 0.031 mm2 using 65 nm CMOS.

Maximum likelihood UEP video streaming technique for WirelessMAN-Advancedsystems

Abstract: In this paper, a maximum likelihood (ML) unequal error protection (UEP) video streaming technique is proposed for WirelessMAN-Advanced systems that are equipped with multiple antennas. The assumed video compression method is H.264/AVC which is a commercially successful technique. ML signal detection is optimal in terms of packet error performance for multiple antenna systems. However, common wisdom is that the optimal ML detection cannot be used for an implementation of UEP which is important for video streaming. In this paper, it is shown that UEP can be implemented using the optimal ML detection. With the aid of computer simulations, the proposed ML-UEP technique is shown to improve video quality significantly compared to a recently proposed UEP technique that is based on the suboptimal signal detection QR-LRL. The PSNR improvement by the proposed method comes at the cost of increased computational complexity in the physical layer 1.

Unequal error protection technique for video streaming over MIMO-OFDM systems

Abstract: In this paper, a novel unequal error protection (UEP) technique is proposed for video streaming over multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) systems. Based on the concept of hierarchical quadrature amplitude modulation (HQAM) UEP and multi-antenna UEP, the proposed technique combines the relative protection levels (PLs) of constellation symbols and the differentiated PLs of the transmit antennas. In the proposed technique, standard square quadrature amplitude modulation (QAM) constellations are used instead of HQAM so that the QAM mapper at the transmitter side and the soft decision calculation at the receiver side remain unchanged, but the UEP benefit of HQAM is retained. The superior performance of the proposed technique is explained by the improved connections between data with various priorities and data paths with various PLs. The assumed video compression method is H.264/AVC, which is known to be commercially successful. The IEEE802.16m system is adopted as a data transmission system. With the aid of realistic simulations in strict accordance with the standards of IEEE802.16m systems and H.264/AVC video compression systems, the proposed technique HQAM-multi-antenna UEP is shown to improve the video quality significantly for a given average bit error rate when compared with previous techniques.