Optimal training design and channel estimation for spatially correlated multi-user multi-input multi-output with orthogonal frequency-division multiplexing (MIMO-OFDM) systems is still an open research topic of great interest. This study first applies tractable semi-definite programming (SDP) to obtain the optimal training signal for the general case of spatial channel correlations for multi-user MIMO-OFDM. In order to reduce the computational complexity of the SDP-based solution, an approximate solution in closed-form is then presented. For a special case of transmit correlations, an optimal solution in closed-form expression is also derived. Analytical and simulation results demonstrate the excellent performance of the proposed designs and their performance advantage over the existing equi-powered training designs.

Optimal training design and channel estimation for spatially correlated multi-user multi-input multi-output with orthogonal frequency-division multiplexing (MIMO-OFDM) systems is still an open research topic of great interest. This paper applies tractable semi-definite programming to obtain the optimal training signal for the general case of spatial channel correlations for multi-user MIMO-OFDM. The optimal solution in closed-form expression is also derived for a special case of transmit correlations. Analytical and simulation results demonstrate the excellent performance of the proposed designs and their performance advantage over the equi-powered training designs.

Widespread deployment of Digital Subscriber Line (DSL) Internet service is impeded by faults in the telephone lines and, for high-speed video services, even a minor discontinuity is important. Identification of faults can help service providers "pre- qualify" lines for DSL service and fault monitoring is essential to maintain an established service. Some faults affect many customers and should be repaired immediately while others, affecting only one customer, degrade slowly and repairs can be scheduled later. Thus, repair can be more effectively scheduled if the type of fault is known. With sufficiently sensitive test equipment, the entire length of the telephone line could be tested from the switching center which avoids costly dispatch of technicians and facilitates automated monitoring. Our method uses frequency domain reflectometry (FDR) and measurements show extended range with enhanced sensitivity and resolution when compared to the well- known time domain reflectometer (TDR). The FDR method easily detects a change of wire gauge at 1200 meters while a commercial TDR instrument is barely able to do so. To remotely identify the fault type, we introduce a new measurement of reflection phase angle: water in the cable (~160deg), bad splice (~10deg), gauge changes (~135deg), bridged taps (180deg), and open circuits (0deg).

This paper is concerned with the bit error probability (BEP) of coded unitary space-time modulation systems based on finite-length low density parity check (LDPC) codes. The union bound on the BEP is derived for any code rate, unitary space-time constellation and mapping. The tightness of the bound is verified with simulation results of the ordered statistic decoding (OSD). Numerical and simulation results show that the union bound is also close to the error performance of the sum-product decoding at the low bit error probability (BEP) levels when Gray mapping is employed. The derived bound is useful to benchmark the error performance of LDPC coded unitary space-time modulation systems.

This paper reviews established methods of communication using in-building power conductors with particular attention to the 100 kHz to 450 kHz frequency range. It also reviews line characteristic impedance and it presents new measurements of attenuation and noise on the power conductors. Measurements illustrate interfering noise caused by individual electrical loads and also the combined noise at power outlets in an office environment. Using a simple FSK transmission system, bit errors were measured as a function of time and these were found to coincide with noise bursts that occur at regular times during the 60 Hz cycle. The paper concludes by proposing an interference avoidance scheme where transmission is suspended during periods of high interference and then continues during quiet intervals.

This paper proposes a two-phase hybrid algorithm for optimal placement of radioports. The placement strategy minimizes the dynamic range of the channel, and thus reduces the peak transmit power needed by a wireless handset. The performance of the proposed algorithm is compared with that of an exhaustive search, as well as with that of a simplex algorithm. Simulation results show that the two-phase hybrid algorithm, although suboptimal, outperforms the simplex algorithm with only a modest increase in computation time. It has the added advantage of being quite insensitive to the seed solution.

This paper presents a segmented matched filter (SMF) for codephase acquisition in direct sequence spread spectrum systems. While conventional matched filters provide fast acquisition in the presence of high co-user noise, they are unable to handle significant carrier frequency offset (Doppler). This problem is alleviated by segmentation with non-coherent summation. The paper develops expressions to approximately relate the matched filter partitioning to the pre-detection filter and dwell time integrator of the conventional non-coherent correlator. It also investigates 1-bit versus 2-bit quantization. A mixed-signal application specific integrated circuit (ASIC) has been fabricated to implement a 512 chip SMF with half chip codephase resolution. The paper presents calculated and measured probability density functions (pdf) for the filter output decision variable for 10, 25, and 50 co-users with 0 to 20 kHz Doppler shift. For the example of a GPS receiver, expected acquisition time is shown as a function of multiple access interference and carrier Doppler shift.

A simple negative impedance amplifier has been developed that can bidirectionally amplify a wideband signal on a telephone line. The amplifier uses a multibranch RC network to accurately match the negative impedance to the characteristic impedance of the line over a wide range of frequencies. An operating range extension of 1 km has been achieved which gives an increase of 40% in area coverage. Advantages of this amplifier are low complexity and good tolerance to component variation. A 2 wire/4 wire amplifier has also been developed. It has achieved a range increase of 1.5 km which is 27% more range or 61% more area coverage. The amplifier can also operate closer to either end of the line than a negative impedance amplifier of the same gain. A disadvantage of the 2w/4w amplifier is the increase in circuit complexity by approximately a factor of four. These techniques are not dependent on the modulation scheme and thus a single unit could be used in ADSL, HDSL, or BR-ISDN loops.

A new Z search method is proposed which can significantly improve mean acquisition time when compared with the best previous Z search strategy. The codephase search starts from a position offset from centre of a range which has been determined from a priori information. The proposed search method includes previous Z search strategies as special cases. The distribution and mean of acquisition time have been developed for the proposed search strategies. With minimum mean acquisition time as objective, optimization of the starting position offset has been investigated.

The paper presents a time-domain approach to the analysis of the code phase acquisition process in direct sequence spread spectrum receivers. This approach is suitable for a variety of search and verification/lock strategies. The time-domain approach is more direct than the unified z transform-domain approach and permits computation of the acquisition time distribution. The paper also provides equations for calculating the mean acquisition time under various search and verification/lock strategies. The usefulness of the approach is demonstrated by application to a specific nontrivial case of practical significance.

Serial search spread-spectrum code acquisition is analyzed. A modified flow graph is presented. It permits the use of a generalized lock strategy and allows the false alarm time to be treated as a random variable. The distribution of acquisition time is obtained directly by using an extended generalization of Bernoulli trials. Compared to recent work which uses a fixed penalty false alarm time, the analysis is more general and requires fewer assumptions and approximations. The analysis is suitable for both single dwell and multidwell systems. Examples are presented to show the effect on the acquisition time distribution of sequence length and normalized detection threshold.>

The authors evaluate a robust framing technique which incorporates sufficient memory to simultaneously evaluate all bits in a candidate synchronizing codeword. Three performance measures are identified and the distribution function and mean value are developed in closed form for each measure. Predicted search and maintenance time distributions have been verified by measurements on a laboratory model. The authors present graphs of search time distributions which illustrate the effect of system parameters and variations in the channel error rate.>

Serial digital transmission requires reliable frame synchronization at the receiver in order to demultiplex and recover the transmitted signals. This paper proposes a faster and more robust method of detecting frame synchronization. A large serial memory is used to store sufficient received history to include all bits of the synchronizing codeword. Equations have been developed which predict the search and "in frame" performance. These equations have been verified with laboratory models. Using, for example, the well-known D3 PCM system, the reframe time is improved from 48 to 3 ms. With an error rate of 1 percent, the predicted "in frame" time is improved from 1.9 s to several centuries! Techniques for hardware implementation are suggested and application to the extended framing format is discussed.

This paper investigates the adaption of qnantizer step size in voice coding circuits. Techniques which maintain the decoder step size in the presence of channel errors have been evaluated. Graphs are presented for the comparison of error recovery performance. Practical considerations favor one algorithm, and its digital implementation has been described in the paper. Laboratory measurements of step-size error recovery and SNR performance have been included to verify the theory.