Resource Allocation of MU-OFDM Based Cognitive Radio Systems Under Partial Channel State Information

Chunyan MiaoarXiv 2008 · arXiv

TL;DR

This paper tackles the challenge of efficiently allocating resources in cognitive radio systems using MU-OFDM when the transmitter has imperfect knowledge of the wireless channel. They propose a novel two-step resource allocation strategy that balances maximizing data rates with ensuring quality of service (QoS) under these realistic channel conditions.

What Problem Does This Paper Address?

In wireless communication systems, especially those employing adaptive transmission techniques, it's often assumed that the transmitter has perfect knowledge of the channel state information (CSI). However, in real-world wireless scenarios, this assumption is flawed due to factors like feedback delays, estimation errors, and quantization. When the transmitter operates with incomplete CSI, the system's performance can degrade, potentially failing to meet Quality of Service (QoS) requirements. Cognitive Radio (CR) systems, which aim to improve spectrum efficiency by allowing secondary users (CRUs) to utilize unused licensed bands, are particularly sensitive to these imperfections. CRUs must adapt their transmission parameters rapidly to avoid interfering with primary users (PUs). Orthogonal Frequency Division Multiplexing (OFDM) is a promising modulation scheme for CR systems due to its flexibility. However, optimally allocating subcarriers, bits, and transmit powers in a multiuser OFDM (MU-OFDM) CR system with imperfect CSI is a complex problem that has not been extensively addressed. This paper specifically investigates this challenge for downlink MU-OFDM CR systems.

Methods

The authors propose a two-step approach to address the constrained multiuser resource allocation problem:

  1. Subcarrier Allocation: They introduce a simple algorithm called SAMA (Subcarrier Allocation with MA) which aims to reduce the computational complexity of this step. SAMA involves setting a threshold to discard less favorable subcarriers and then allocating the remaining ones based on a calculated channel factor.
  2. Transmit Powers and Bits Allocation: For this step, they employ a Memetic Algorithm (MA). MAs are a hybrid approach combining genetic algorithms with local search methods, known for their efficiency in solving complex combinatorial optimization problems. This MA is used to determine the optimal bits allocation to CRUs after subcarriers have been assigned.

The paper also addresses the challenge of modeling the imperfect CSI. They assume that the CSI is obtained at the receiver perfectly and then fed back to the transmitter with a delay. The transmitter models the channel using a correlation coefficient derived from Jake's model, accounting for the imperfect feedback. To handle the complex bit error rate (BER) function under partial CSI, they approximate it using a Nakagami-m distribution.

Datasets

Not discussed in the paper. The authors use simulated data for their experiments.

Key Insights and Ideas

  • Partial CSI is a significant performance bottleneck: The simulations clearly demonstrate that the degree of correlation between the estimated CSI and the actual CSI (represented by 'p') has a substantial impact on the total CRU bit rate. Even small decreases in 'p' lead to significant performance degradation.
  • User diversity matters: When the data rate requirements for different CRUs become less uniform (i.e., some users have much higher requirements than others), the overall system performance decreases. This highlights the importance of user diversity in achieving higher aggregate throughput.
  • The proposed two-step approach is effective: The combination of SAMA for subcarrier allocation and MA for bit allocation provides a computationally tractable solution to a complex optimization problem.
  • Trade-offs exist: The system needs to balance maximizing data rates with satisfying QoS constraints (like BER) and managing interference to primary users.

Key Results

The paper presents simulation results that illustrate the impact of various parameters on the average total CRU bit rate ($R_s$). Key findings include:

  • Impact of Partial CSI (p): As 'p' decreases (indicating less accurate CSI), the total CRU bit rate ($R_s$) consistently decreases across different data rate requirement scenarios ($\lambda$). For instance, with $\lambda = [1111]$, the bit rate for $p=1$ is always higher than for $p=0.9$ and $p=0.7$.
  • Impact of Data Rate Requirements ($\lambda$): When the data rate requirements become less uniform (e.g., $\lambda = [1118]$ compared to $\lambda = [1111]$), the total bit rate ($R_s$) generally decreases. This is attributed to a decrease in user diversity.
  • Impact of Power Budget ($P_{total}$): The total CRU bit rate ($R_s$) increases with the maximum transmit power budget ($P_{total}$), as expected.
  • Impact of Interference Power ($I_{total}$): The total CRU bit rate ($R_s$) also increases with the tolerable interference power ($I_{total}$), though the sensitivity to $I_{total}$ appears to be less pronounced than to $P_{total}$ in some scenarios.

Limitations and Future Work

  • Approximation of BER: The paper uses a Nakagami-m distribution to approximate the BER function. While this simplifies the problem, it introduces an approximation that might affect the exactness of the solution.
  • Focus on Downlink: The current work focuses on downlink transmission. Future work could extend this to uplink scenarios.
  • Simplified CSI Model: The CSI model assumes perfect CSI at the receiver and only considers feedback delay and correlation. More complex CSI imperfections (e.g., quantization errors at the receiver) could be investigated.
  • PU Interference Modeling: While interference to PUs is considered, more sophisticated models for PU behavior and interference management could be explored.
  • Algorithm Optimization: Further research could explore more advanced optimization algorithms for both subcarrier and bit allocation to potentially achieve even better performance.

Summary updated

Abstract

This paper has been withdrawn by the author due to some errors.
BibTeX
@misc{miao2008resource,
  author = {Chunyan Miao},
  title = {{Resource Allocation of MU-OFDM Based Cognitive Radio Systems Under Partial Channel State Information}},
  year = {2008},
  eprint = {0808.0549},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/0808.0549},
}
Figure