Transmit without regrets: Online optimization in MIMO-OFDM cognitive radio systems

Panayotis Mertikopoulos, Elena Veronica Belmega

Abstract

In modern wireless communications, cognitive radio systems must evolve dynamically to accommodate changing user demands and environmental conditions. This paper investigates a MIMO-OFDM cognitive radio network where secondary users (SUs) must adapt their transmit profiles in real-time to maximize data rates without interfering with primary users (PUs). Because the environment is highly volatile, traditional static solution concepts like Nash equilibrium are often inapplicable. We propose a dynamic transmit policy based on "no-regret" optimization. By utilizing Augmented Exponential Learning (AXL) and online mirror descent techniques, we derive a policy that relies only on local channel state information (CSI). Our results demonstrate that this approach allows SUs to track optimal transmit profiles effectively, even under rapid environmental changes and significant observation errors.

Introduction

The radio spectrum is a finite and increasingly crowded resource. Cognitive Radio (CR) technology addresses this scarcity by allowing Secondary Users (SUs) to utilize frequency bands that are temporarily vacated by licensed Primary Users (PUs). However, the practical implementation of CR is complicated by the inherent volatility of the wireless medium. PUs arrive and depart unpredictably, and signal quality fluctuates due to physical movement and fading.

In such a dynamic setting, the goal for SUs is to optimize their transmission profiles—specifically, how power is distributed across frequency subcarriers (OFDM) and spatial directions (MIMO). Traditional optimization methods often assume a static or slowly changing environment where users can converge to a Nash Equilibrium. In reality, by the time an equilibrium is calculated, the network state has often shifted. This paper shifts the focus toward "no-regret" policies, which ensure that a user's average performance is at least as good as the best fixed strategy they could have chosen in hindsight.

The Augmented Exponential Learning (AXL) Framework

To address the challenges of dynamic MIMO-OFDM systems, we introduce the Augmented Exponential Learning (AXL) algorithm. This framework is rooted in online mirror descent and matrix exponential learning. The algorithm decomposes the complex optimization task into two primary components:

  1. Power Allocation (OFDM): Determining the optimal distribution of power across various frequency subcarriers.
  2. Signal Covariance (MIMO): Managing the spatial direction of signals across multiple antennas to maximize throughput.

AXL assigns "scores" to different transmission strategies based on their historical performance, updating these scores exponentially. This method offers several critical advantages for real-world deployment:

  • Distributed Operation: Each SU requires only local data, eliminating the need for a central controller.
  • Statelessness: Users do not need to know the global network topology or the specific strategies of other users.
  • Robustness: The algorithm is mathematically proven to maintain performance even when Channel State Information (CSI) measurements are noisy or imperfect.

Experimental Methodology

The performance of the AXL algorithm was evaluated through extensive numerical simulations designed to mimic realistic wireless environments.

  • Channel Modeling: We employed the Jakes model to simulate Rayleigh fading, representing signal fluctuations typical of pedestrian speeds (3–5 km/h).
  • Network Configuration: The simulation environment consisted of 10 Primary Users and 40 Secondary Users communicating across 256 subcarriers.
  • Traffic Dynamics: PU activity was modeled using a Poisson process to create a stochastic, "bursty" environment that tests the agility of the secondary users.

Key Results and Discussion

The study yielded several significant insights into the efficacy of no-regret learning in cognitive radio:

Convergence and Regret Bounds

We proved that AXL achieves a worst-case average regret of $O(T^{-1/2})$ over $T$ time steps. This is a theoretically tight bound, representing the best possible performance guarantee in such a volatile environment. In practice, AXL reached maximum system capacity within just 3–4 iterations.

Comparison with Traditional Methods

When compared against the standard Simultaneous Water-Filling (SWF) algorithm, AXL demonstrated superior stability. While SWF often failed to converge in dynamic scenarios—leading to oscillations and suboptimal throughput—AXL consistently tracked the evolving optimum.

Robustness to Noise

A defining feature of AXL is its resilience to measurement errors. In simulations where 50% relative error was introduced into channel measurements, AXL still achieved system capacity. While the convergence was slower than in the perfect CSI case, traditional methods failed entirely under these conditions, offering no improvement over random initialization.

Conclusion

The transition from equilibrium-based strategies to no-regret learning represents a significant step forward for cognitive radio systems. The AXL algorithm provides a robust, distributed, and efficient means for secondary users to navigate the complexities of MIMO-OFDM networks. While future work may explore "adaptive regret" for even higher mobility scenarios (such as high-speed rail) or incorporate energy-aware constraints, the current findings establish AXL as a powerful tool for smart spectrum sharing in a moving world.

Summary updated

Abstract

In this paper, we examine cognitive radio systems that evolve dynamically over time due to changing user and environmental conditions. To combine the advantages of orthogonal frequency division multiplexing (OFDM) and multiple-input, multiple-output (MIMO) technologies, we consider a MIMO-OFDM cognitive radio network where wireless users with multiple antennas communicate over several non-interfering frequency bands. As the network's primary users (PUs) come and go in the system, the communication environment changes constantly (and, in many cases, randomly). Accordingly, the network's unlicensed, secondary users (SUs) must adapt their transmit profiles "on the fly" in order to maximize their data rate in a rapidly evolving environment over which they have no control. In this dynamic setting, static solution concepts (such as Nash equilibrium) are no longer relevant, so we focus on dynamic transmit policies that lead to no regret: specifically, we consider policies that perform at least as well as (and typically outperform) even the best fixed transmit profile in hindsight. Drawing on the method of matrix exponential learning and online mirror descent techniques, we derive a no-regret transmit policy for the system's SUs which relies only on local channel state information (CSI). Using this method, the system's SUs are able to track their individually evolving optimum transmit profiles remarkably well, even under rapidly (and randomly) changing conditions. Importantly, the proposed augmented exponential learning (AXL) policy leads to no regret even if the SUs' channel measurements are subject to arbitrarily large observation errors (the imperfect CSI case), thus ensuring the method's robustness in the presence of uncertainties.
BibTeX
@article{mertikopoulos2014transmit,
  author = {Panayotis Mertikopoulos and Elena Veronica Belmega},
  title = {{Transmit without regrets: Online optimization in MIMO-OFDM cognitive radio systems}},
  journal = {IEEE J. Sel. Areas Commun.},
  year = {2014},
  url = {https://openreview.net/forum?id=ol5A7xctdiT},
}
Figure