Towards the Asymptotic Sum Capacity of the MIMO Cellular Two-Way Relay Channel

Xiaojun Yuan, Zhaoxi Fang, Xiaojun Yuan 0002, Xin WangIEEE TSP 2014 · DOI

TL;DR

This paper introduces a novel two-way relaying scheme for cellular networks that significantly improves spectral efficiency. By employing a new lattice precoding technique, the proposed method effectively combats inter-stream interference, allowing it to approach the theoretical sum capacity of MIMO cellular two-way relay channels, especially in high signal-to-noise ratio (SNR) scenarios.

What Problem Does This Paper Address?

Cellular networks are increasingly relying on two-way relaying to enhance spectral efficiency, allowing multiple mobile stations (MSs) to exchange information with a base station (BS) via a relay station (RS). Existing methods for MIMO cellular two-way relay channels (cTWRC) often struggle with inter-stream interference (ISI), particularly when MSs cannot cooperate. This limitation prevents them from achieving the theoretical maximum data rate (sum capacity). Previous approaches, like amplify-and-forward (AF) or decode-and-forward (DF) with linear precoding, either suffer from noise amplification or fall short of the capacity due to insufficient interference management, especially in complex MIMO setups. This paper aims to bridge this gap by proposing a scheme that can asymptotically achieve the sum capacity.

Methods

The core innovation of this work lies in a new non-linear lattice precoding technique applied at the BS during the first phase of transmission. This technique, combined with linear precoding, successive interference cancellation (SIC) at the relay, and nested lattice coding, effectively pre-compensates for inter-stream interference. This allows for interference-free lattice decoding at the relay, a crucial step for efficient two-way communication.

Key methodological aspects include:

  • Lattice Precoding: A novel non-linear precoding method to mitigate ISI before it reaches the relay.
  • Nested Lattice Coding: Utilized for its ability to achieve high rates and its compatibility with interference cancellation.
  • Successive Interference Cancellation (SIC): Applied at the relay to sequentially remove interference, enabling the decoding of individual data streams.
  • Channel Triangularization: Using QR and RQ decompositions to simplify the channel matrices, facilitating the analysis and design of the precoding and decoding schemes.
  • Monotonic Programming (MP): The problem of optimizing power allocation for weighted sum-rate maximization is formulated as a monotonic program, which is then solved using an efficient polyblock outer approximation algorithm for guaranteed global optimality.

Datasets

Not discussed in the paper. The evaluation is based on simulated channel models.

Key Insights and Ideas

  • Lattice Precoding for Interference Management: The paper demonstrates that non-linear lattice precoding can be a powerful tool to manage inter-stream interference in MIMO cTWRC, a problem that linear precoding alone struggles to solve effectively.
  • Asymptotic Capacity Achievement: The proposed scheme is shown to asymptotically achieve the sum capacity of the MIMO cTWRC under certain conditions, particularly in high SNR regimes. This is a significant improvement over prior work.
  • Power Allocation Optimization: The formulation of the weighted sum-rate maximization problem as a monotonic program and its solution via a polyblock outer approximation algorithm provide a robust method for optimizing system performance across various SNR conditions.
  • Flexibility in DPC Order: The analysis highlights that the performance of the proposed scheme can be influenced by the DPC (dirty-paper coding) re-encoding order at the relay, suggesting that optimal selection of this order is crucial for maximizing performance.

Key Results

  • The proposed scheme significantly outperforms existing schemes, including AF-IA and Yang's scheme, in terms of achievable sum-rate.
  • In high SNR regimes, the proposed scheme closely approaches the theoretical cut-set bound, with gaps as small as 0.5 bps/Hz.
  • For a cTWRC with K=4 MSs, the proposed scheme offers a gain of about 3 bps/Hz over Yang's scheme.
  • With a larger number of MSs (K=16), the proposed scheme provides a power gain of over 7 dB compared to Yang's scheme.
  • The optimal power allocation algorithm ensures that the system can achieve near-capacity performance even when the network is not bottlenecked by a specific link.

Limitations and Future Work

  • Computational Complexity: While the proposed scheme offers significant performance gains, finding the optimal DPC encoding order can be computationally intensive, especially for a large number of MSs. The paper mentions using random searches for suboptimal orders.
  • General Antenna Setups: The paper discusses extensions to general antenna setups and multiple relays, suggesting that further research in these areas could enhance performance.
  • Channel Estimation: The current work assumes perfect CSI is available. Practical implementation would require robust channel estimation techniques, which is acknowledged as a challenging but potentially affordable aspect given the infrastructure.
  • Cooperative MSs: The current scheme assumes MSs cannot cooperate. Investigating scenarios where MSs can cooperate could lead to further performance improvements.

Summary updated

BibTeX
@article{yuan2014towards,
  author = {Xiaojun Yuan and Zhaoxi Fang and Xiaojun Yuan and Xin Wang},
  title = {{Towards the Asymptotic Sum Capacity of the MIMO Cellular Two-Way Relay Channel}},
  journal = {IEEE TSP},
  year = {2014},
  doi = {10.1109/tsp.2014.2332971},
}
Figure