Yao-Nan Lee

A novel indoor RSS-based position location algorithm using factor graphs

A low-complexity and accurate receive-signal strength (RSS)-based algorithm is proposed to estimate a target location. To achieve the aim of low complexity, a local linearization technique based on the local linearity in the surface of power decay profile (PDP), established by training logarithmic RSS measurements and collected from each individual access point (AP), was devised. To achieve a near-optimum solution, the stochastic properties of measurement errors and the reliability of the measurement data are introduced into the factor graph framework. Numerical experiments show that the proposed algorithm not only achieves a near maximum likelihood (ML) solution based on training RSS measurements, but also enjoys low complexity.

Normal Graphs for Downlink Multiuser MIMO Scheduling

Inspired by the success of the low-density parity-check (LDPC) codes in the field of error-control coding, in this paper we propose transforming the downlink multiuser multiple-input multiple-output scheduling problem into an LDPC-like problem using the normal graph. Based on the normal graph framework, soft information, which indicates the probability that each user will be scheduled to transmit packets at the access point through a specified angle-frequency sub-channel, is exchanged among the local processors to iteratively optimize the multiuser transmission schedule. Computer simulations show that the proposed algorithm can efficiently schedule simultaneous multiuser transmission which then increases the overall channel utilization and reduces the average packet delay.

Asymptotic spectral efficiency of MIMO multiple-access wireless systems exploring only channel spatial correlations

We use the replica method originally developed in statistical physics to investigate the asymptotic sum-rate of a Gaussian antenna-array-based multiple-input multiple-output (MIMO) multiple-access wireless channel having spatial correlations at both the transmitters and the receiver. The asymptotic solution is not only rigorously valid for systems with large array sizes, hut it also produces highly accurate ergodic results for systems with only a few antenna elements at each transmitter and receiver. This otters the asymptotic solution important practical values in analyzing and designing a MIMO multiple access system that makes best use of the wireless channel structure. Furthermore, with the asymptotic solution, we provide an efficient iterative water-filling algorithm to determine the optimum transmit signal covariance matrices when only the slow-varying channel spatial covariance information is available.

Performance analysis of MIMO wireless with spatially-correlated channels. Part II. Separate-decoding

For pt.I, see ibid., p.3504-8 (2004). In Part I of this paper, we analyzed the asymptotic mutual information of the joint-decoding MIMO systems. Here in Part II, we focus on the performance analysis of the MIMO systems with various separate decoding schemes, i.e., with various suboptimum spatial equalizers followed by a bank of temporal decoders. Among the suboptimum spatial equalizers, under the proposed framework, we cover both 1) linear spatial equalizers such as linear MMSE spatial equalizer, linear zero-forcing spatial equalizer, and matched-filter equalizer, and 2) nonlinear spatial equalizers such as individually optimal spatial equalizer and marginal optimal spatial equalizer. Closed-form solutions are derived for all the possible combinations of 1) various transmit-side arbitrary input distributions and 2) the receive-side decoding schemes listed above. Important physical meanings revealed though the closed-form solutions to those more practical combinations, as how the input-output mutual information is affected by the transmit-side and the receive-side channel spatial correlations, are highlighted along our derivation.

CTH09-2: A Time-Efficient Approach for Designing LDPC-Coded MIMO Systems

With extrinsic information transfer (EXIT) chart code design tool, people can more easily design a capacity-approaching low-density parity-check (LDPC) codes for antenna-array-based multiple-input multiple-output (MIMO) systems. However, extremely time-consuming Monte-Carlo simulations are required to evaluate the statistics used in the EXIT charts, which makes capacity-approaching LDPC code design almost impossible for large-scale MIMO systems. By making use of large-system performance analysis technique, we propose a time-efficient code design approach which eliminates the necessity of time-consuming Monte Carlo simulations and then makes capacity-approaching LDPC code design possible for large-scale MIMO systems. Numerical experiments show that the so-constructed LDPC codes attain comparable bit error rate performance with existing well-designed LDPC codes. In addition, we demonstrate that redesigning LDPC code is very helpful only when the ratio of number of transmit and that of receive antennas of a MIMO system varies.

Design of LDPC-coded MIMO systems via a large-system approach

Making use of the extrinsic information transfer (EXIT) chart code design tool and the large-system performance analysis technique, we develop an efficient approach to construct low-density parity-check (LDPC) codes for multiple-input multiple-output (MIMO) systems. In the proposed approach, we eliminate the time-consuming procedure of numerically evaluating the EXIT curves of the MIMO detectors, which makes capacity-achieving LDPC code design possible for large-scale MIMO systems. In addition, we demonstrate that redesigning LDPC code is very helpful only when the I/O ratio of a MIMO system varies.

Spectrum efficiency of MIMO multiple-access wireless systems exploring only channel spatial correlations: an asymptotic approach

We use the replica method originally developed in statistical physics to investigate the asymptotic sum-rate of a Gaussian antenna-array-based multiple-input multiple-output (MIMO) multiple-access wireless channel having spatial correlations at both the transmitters and the receiver. The asymptotic solution is not only rigorously valid for systems with large array sizes, but it also produces highly accurate ergodic results for systems with only a few antenna elements at each transmitter and receiver. Furthermore, with the asymptotic solution, we provide an efficient iterative water-filling algorithm to determine the optimum transmit signal covariance matrices when only the slow-varying channel spatial covariance information is available.