Alexander Seeger

Incremental redundancy and bit-mapping techniques for high speed downlink packet access

A major evolution of the UMTS standard is high speed downlink packet access (HSDPA), which provides higher cell throughput and peak data rates of 10.8 Mbps. Key enabling technologies include fast scheduling, adaptive modulation and coding (AMC), as well as hybrid automatic repeat request (HARQ). To ease the implementation of HSDPA all coding elements are identical to standard WCDMA. In particular, the rate 1/3 turbo code is reused and no trellis coded modulation scheme has been introduced. The resulting drawbacks are analyzed and it is shown how they are efficiently counteracted by bit-mapping techniques accounting for the different bit reliabilities within a 16-QAM symbol. This paper focuses on the implementation and performance of HARQ and bit-to-symbol mapping functions. The performance of different HARQ schemes is compared for QPSK and 16-QAM and varying code rates. It is shown that depending on the code rate and the user terminal capabilities different HARQ types are favorable.

Variable orthogonality factor: a simple interface between link and system level simulation for high speed downlink packet access

Detailed simulations of the full physical layer operation of a cellular network are usually performed in two stages, link level and system level. Emergence of new link adaptation and radio resource management techniques together with need for higher modeling accuracy are blurring the line separating these stages. In this paper we address the issue of performing a link level/system level separation and defining the interface between them for simulating high speed downlink packet access (HSDPA), a part of WCDMA standard. The main challenge is to model frequency selective fading on the system level, while retaining the Rake receiver on the link level. We present available strategies and propose a novel interface based on variable orthogonality factor (OF), which serves as a compact description of the momentary channel impulse response. Finally, we present HSDPA simulation results demonstrating the high accuracy of our approach.

Downlink eigenbeamformer with combining of eigenbeams

For W-CDMA several closed-loop Tx diversity concepts employing 4 antenna elements at the base station (BS) are currently discussed to be included in 3GPP release 5. Among these, the downlink eigenbeamformer effectively uses long-term as well as short-term channel properties. More specifically, it is based on an eigenanalysis of the long-term spatial covariance matrix at the mobile station (MS). The eigenvectors (vector of complex weighting factors for all antenna elements) with the largest eigenvalues (largest average signal power) are determined and fed back step by step to the BS. This process takes place on the same time scale as the MS physical movement. Accordingly, required operations in the MS for eigenanalysis as well as required long-term feedback bits for representation of eigenvectors are distributed over a very large number of transmission slots. As an extension of simple switching between eigenbeams a linear combination of several eigenbeams is considered. This combining is performed using constellation rotation and progressive refinement techniques applied to the eigenbeam space. Performance of combining methods is compared with an upper bound (beamforming based on perfect short-term knowledge of the downlink channel) and uplink based beamforming methods where only the long-term spatial covariance matrix is known at the BS. Although combining methods require more short-term feedback information then simple switching and thus are more sensitive to feedback delay and restricted feedback rate, they prove to be beneficial up to a vehicle speed of 40 km/h.

Antenna weight verification for closed loop transmit diversity

Closed loop transmit diversity (CLTD) for FDD WCDMA relies on low-rate feedback to achieve both beamforming and diversity gain. Since the feedback channel is not immune to errors, occasionally base station (BS) uses different antenna weight vector from the one requested by mobile station (MS). Surprisingly, most of the resulting performance degradation is caused not by reduced power of the Rx signal, but by erroneous dedicated channel estimation at the MS relying on knowledge of used weight vector. In this paper we introduce a general trellis-based antenna weight verification algorithm, which attempts to detect feedback errors and determine the most likely weight vector. By exploiting the structure of CLTD schemes proposed for WCDMA with 2 and 4 BS-antennas we show methods to reduce the complexity of our algorithm. Our comparative simulations show that antenna weight verification reduces the performance degradation due to feedback errors less than 0.5 dB at 1% frame error rate for considered 4 antenna schemes.

Antenna weight verification for closed-loop downlink eigenbeamforming

Adaptive antennas at the base stations have a big potential to increase downlink capacity and coverage in WCDMA systems. One of the most promising techniques of adaptive antenna control is closed-loop transmit diversity. This method achieves beamforming gain as well as diversity gain by feedback of downlink fast fading characteristics from mobile station to base station. Due to the limited feedback rate, typically the best antenna weight vector chosen from a restricted set is reported. If this feedback is subject to transmission errors, two performance degrading effects occur: a suboptimal weight vector is used and channel estimation errors take place at the mobile station. The latter effect arises because the mobile station derives the channel estimate for the dedicated channel from the reported weight vector and the channel estimate per antenna based on the common pilot channel. In effect, feedback errors severely distort this estimate and lead to an error floor effect. However, channel estimation can also be based on the pilot symbols in the dedicated channel itself. This estimate has high variance, but is not distorted by feedback errors. If both estimates are combined in a process called antenna weight verification, performance can be dramatically increased. Within this paper an antenna weight verification scheme for eigenbeamformer is presented, which maximizes the estimation accuracy and eliminates the error floor. This is demonstrated by theoretical analysis and link-level simulations. It is shown that this antenna weight verification reduces the performance degradation to 0.2 dB at 1% frame error rate.

Performance of downlink eigenbeamformer with realistic feedback transmission

Several closed-loop transmitter diversity concepts for multiple antennas are currently under discussion for inclusion in the 3GPP W-CDMA standard. The focus is to find a concept using 4 transmitter antennas allowing for finer beamsteering while still being compatible with the current uplink feedback rate of one bit per transmission slot. Among these concepts, the downlink eigenbeamformer effectively uses long-term as well as short-term channel properties. More specifically, it is based on an eigenanalysis of the long-term spatial covariance matrix. The eigenbeams with the largest eigenvalues (largest average SNR) are determined and fed back step by step to the base station (BS). This process takes place on the same time scale as the mobile stations (MS) physical movement. Accordingly, required operations in the MS as well as required long-term feedback bits are distributed over a very large number of transmission slots. In addition, a short term selection between the eigenbeams is carried out at the MS to account for fast fading. Performance in terms of coded block error rate for the downlink eigenbeamformer concept is analysed under a set of spatial channel models. Simulation results show the influence of important parameters like, e.g., the amount of common pilot channel power compared to another transmitter diversity proposal that does not exploit the long term channel statistics. The multiplexing formats between long-term and short-term feedback information are discussed and a detailed complexity analysis is given.

Broadcast communication on fading channels using hierarchical coded modulation

Since Covers (1972) pioneering work on broadcast channels, it has been known that a substantial gain in transmission rate can be achieved by hierarchical modulation with respect to time-sharing. Here, this work is extended in two directions. First, Rayleigh and Rice fading channels are included. Second, limitations imposed by the use of standard signal constellations, such as M-QAM, are discussed. These extensions lead to a refined understanding of the broadcast channel structure as well as practical design rules for hierarchical coded modulation.

Coverage and capacity enhancement of multiservice WCDMA cellular systems via serial interference cancellation

The uplink coverage and capacity of CDMA cellular systems with the conventional single user detector receiver are interference limited. Particularly, during the roll-out phase, the coverage of a CDMA system is uplink limited. Hence, using serial interference cancellation (SIC) at the base station is a low cost option to increase the overall performance. Considering the typical quality of service requirements of mixed services, i.e. voice and data, a new hybrid receiver structure for interference cancellation is proposed. In order to perform system level analysis, the calculation of signal-to-interference ratios is extended to the case of multiple service classes. Given this tooling, the optimum powers of the mobile stations are derived as a function of various system parameters and receiver structures. This enables an accurate calculation of the intracell and intercell interference and a derivation of coverage-capacity tradeoff expressions. Results show significant performance gains in terms of user capacity and cell coverage by using SIC receivers including the proposed hybrid structure that meets the delay and complexity requirements of the different service classes.

The coverage-capacity tradeoff in multiservice WCDMA cellular systems with serial interference cancellation

The uplink coverage and capacity of CDMA cellular systems with the conventional single user detector receiver are interference limited. Particularly, during the roll-out phase, the coverage of a CDMA system is uplink limited. Hence, using serial interference cancellation (SIC) at the base station is a low cost option to improve the overall performance. Considering the typical quality of service requirements of mixed services, i.e. voice and data, a new hybrid receiver structure for interference cancellation is proposed. In order to perform system level analysis, the calculation of signal-to-interference ratios is extended to the case of multiple service classes with various SIC receiver structures. Given this tooling, the optimum powers of the mobile stations are derived as a function of various system and design parameters. This enables an accurate calculation of the intracell and intercell interference. Based on this, analytical expressions are derived for the coverage-capacity tradeoff. Results show significant performance gains in terms of user capacity and cell coverage by using SIC receivers including the proposed hybrid structure that meets the delay and complexity requirements of the different service classes.

Improving uplink adaptive antenna algorithms for WCDMA by covariance matrix compensation

This document addresses the estimation of long-term spatial covariance matrices. Many beamforming techniques are based on those matrices. In the past, they often have been assumed to be perfectly known, since long averaging processes can be applied. However, a systematic error is committed when using conventional methods for the estimation. We discuss the occurrence and the performance impact of this error. A simple compensation scheme is introduced which completely removes this error. Simulation results are presented for a multiuser WCDMA environment. For beamforming methods based on maximizing the signal power, gains of more than 2 dB are observed. The impact of the systematic error on SINR maximizing algorithms is minor, so that the improvement of the compensation scheme is slight. The results show that an estimate for the interference covariance is needed in all cases, either for compensating the signal covariance or for SINR based techniques.