Abstract
본 논문에서는 시변 다중 입출력 (multiple-input multiple-output) 방송(broadcast) 채널에서 피드백 양자화와 지연을 고려한 협력 빔형성 (coordinated beamforming: CBF) 시스템을 제안한다. 다중 데이터 스트림을 전송하는 CBF 시스템에 피드백 양자화 기법을 적용하고, 구현 복잡도와 피드백 오버헤드 측면에서 효율적인 CBF 시스템을 제시한다. 또한, 실제적인 무선통신 환경에서 발생하는 피드백 지연에 의한 오류를 최소화하기 위하여 사용자 단말에 선형 채널 예측기를 적용한다. 선형 예측기로 Wiener 필터를 이용하여 피드백 지연시간 후의 미래 채널을 예측하고 이를 토대로 피드백 정보를 생성함으로써 지연된 피드백 정보를 이용하는 CBF 시스템의 성능을 향상시킨다. 모의실험을 통해 다양한 도플러 (Doppler) 주파수의 MIMO 방송 채널에서 양자화와 Wiener 필터를 적용한 CBF 시스템의 향상된 심볼 오율과 합 전송률 성능을 확인한다.
Introduction
In Multi-User Multiple-Input Multiple-Output (MU-MIMO) systems, a base station (BS) communicates with multiple users simultaneously. To manage inter-user interference effectively, the BS requires accurate Channel State Information (CSI). However, practical systems face two significant hurdles: feedback delay and quantization error. Feedback delay is particularly problematic in time-varying environments where the physical channel changes between the time of measurement and the time of signal transmission. Quantization error arises because users must compress CSI into a limited number of bits, introducing noise. When ignored, these factors lead to a "performance floor" where increasing signal power (SNR) no longer improves the data rate.
Proposed Coordinated Beamforming (CBF) Scheme
The authors propose a refined Coordinated Beamforming (CBF) architecture designed for time-varying broadcast channels. Unlike many schemes that focus on a single data stream per user to save overhead, this paper proposes a method to support multiple streams per user using equal power allocation. This approach captures multiplexing gains while avoiding the complex feedback overhead typically associated with optimal "Water-Filling" power loading techniques.
Channel Prediction via Wiener Filtering
To mitigate the "stale information" problem caused by feedback delay, the proposed system applies a linear channel predictor at each user terminal. Specifically, a Wiener filter is used to predict the future channel state at the time the signal will actually be transmitted (after the feedback delay $n_d$). This "look-ahead" approach ensures the BS receives information that is relevant to the moment of transmission.
As shown in the terminal architecture, the system estimates the channel, buffers past states, and utilizes the Wiener filter to generate a predicted channel matrix. This predicted information is then quantized using a codebook-based method that exploits the statistical properties of the channel components to minimize the required feedback bits.
Simulation and Performance Analysis
The study relies on numerical simulations using Jake’s model for Rayleigh fading in time-varying channels. The simulation environment uses a 2GHz carrier frequency, an 8ms feedback delay, and varying Doppler frequencies (20Hz to 60Hz) to simulate different user speeds.
Key results include:
- Sum-Rate Gains: At a Doppler frequency of 20Hz, the proposed system with Wiener prediction nearly matches the performance of a system with "Perfect CSI," whereas conventional systems without prediction see their sum-rates plateau quickly.
- Error Reduction: The Symbol Error Probability (SEP) is significantly lower. For example, at 20dB SNR and 30Hz Doppler, the predicted system maintains a steady decline in error, while the non-predicted system hits an error floor.
- Robustness: Even at high speeds (60Hz Doppler), the Wiener filter provides a noticeable performance boost, though the gap to perfect CSI naturally widens as the channel becomes less predictable.
Conclusion and Future Directions
The proposed CBF scheme demonstrates that in dynamic environments, a noisy prediction of the future is far more valuable than a perfect report of the past. While the Wiener filter requires prior knowledge of the channel's autocorrelation, it provides a significant performance advantage. Future research may explore adaptive predictors (like LMS algorithms) that can learn channel dynamics on the fly without pre-calculated statistics, as well as scaling these techniques for massive MIMO systems with many more users.
Summary updated