Abstract
As wireless services expand and the radio frequency spectrum becomes increasingly congested, efficient spectrum multiplexing has become a critical area of research. This paper presents a novel weak signal detection system for Cognitive Radio (CR) based on the chaotic Duffing oscillator. By utilizing a two-level array detection architecture and the proposed TICTACTOEMIDLET algorithm, the system automates the detection of weak primary signals in high-noise environments, significantly improving the speed and accuracy of spectrum hole identification.
Introduction
The rapid growth of wireless communication has turned the radio frequency spectrum into a scarce resource. Cognitive Radio technology addresses this by allowing secondary users to "borrow" unused spectrum bands, known as spectrum holes. However, the technical challenge lies in reliably detecting very weak primary signals at low Signal-to-Noise Ratios (SNR). Traditional detection methods often struggle in these conditions. While chaotic systems like the Duffing oscillator are extremely sensitive to weak periodic signals, their use has historically been limited by the need for manual state observation. This research proposes an automated, grid-based approach to overcome these limitations.
System Architecture
The proposed system employs a "coarse-to-fine" detection strategy to balance speed and precision. The architecture is divided into two main stages:
- Level 1 (Rough Detection): Wideband filters are used to quickly scan the spectrum and eliminate clearly empty bands.
- Level 2 (Fine Detection): Narrowband filters and adjustable local oscillators ($g_0$) focus on "suspicious" bands. The parameters are tuned to keep the system at the "edge of chaos," where it is most sensitive to external periodic signals.
The TICTACTOEMIDLET Algorithm
The core innovation of this work is the TICTACTOEMIDLET algorithm, designed to automate the transition from qualitative phase-plot observation to quantitative mathematical decision-making. The algorithm divides the oscillator's phase space (velocity vs. position) into an $n \times n$ grid:
- Chaotic State: When no signal is present, the system's trajectory wanders widely, filling a high percentage of grid cells.
- Periodic State: When a weak signal is detected, the system collapses into a stable, ordered loop, leaving many grid cells empty.
By calculating the cell-fill ratio ($\lambda$), the system can mathematically determine the presence of a signal. For rough detection, an $8 \times 8$ grid is used, while a $16 \times 16$ grid provides higher precision during the fine detection phase.
Results and Discussion
The algorithm was evaluated through numerical simulations based on the IEEE 802.22 protocol for Wireless Regional Area Networks. Testing focused on detecting weak signals (such as wireless microphones) at 101.17MHz within the 54MHz to 862MHz range.
Key findings include:
- High Sensitivity: The system successfully detected signals with amplitudes as low as 0.0040.
- Mathematical Thresholds: In the rough detection phase, a cell-fill ratio threshold of $\lambda < 0.98$ accurately indicated a transition to the periodic state, signaling the presence of a primary user.
- Computational Efficiency: By replacing complex Lyapunov exponent calculations with simple cell-counting logic, the algorithm significantly reduces computational overhead, making it more suitable for real-time hardware implementation.
Conclusion
This paper demonstrates that a two-array detection system using adjustable Duffing oscillators can effectively automate weak signal detection for Cognitive Radio. The TICTACTOEMIDLET algorithm provides a robust, computationally efficient method for state transition detection. Future work will focus on testing the system in real-world RF environments characterized by multipath fading and non-Gaussian noise to further validate its practical applicability.
Summary updated