Abstract
The proliferation of ubiquitous computing applications has created a multi-dimensional optimization problem involving conflicting variables such as spectral efficiency, complexity, power consumption, delay, and error probability. To address these challenges, it is necessary to augment the overutilized radio spectrum with new frequency bands, such as the optical spectrum, to off-load traffic. This paper presents an efficient system design using amplitude-coherent (AC) detection to reduce the complexity of optical wireless communication (OWC) systems while improving reliability and spectral efficiency. By employing amplitude shift keying (ASK) with orthogonal frequency division multiplexing (OFDM) based on the discrete cosine transform (DCT), the system achieves significant gains. The proposed architecture includes a low-complexity detector, a blind channel estimator, and a peak-to-average power ratio (PAPR) reduction scheme that requires no feedback overhead.
Introduction
As the demand for high-speed wireless connectivity grows, the radio frequency (RF) spectrum has become increasingly congested. Optical Wireless Communication (OWC), which includes Visible Light Communication (VLC) and Free Space Optics (FSO), has emerged as a vital technology to offload traffic. However, traditional OWC systems often face trade-offs between spectral efficiency and hardware complexity. Standard OFDM implementations, for instance, require Hermitian symmetry to produce real-valued signals for intensity modulation, effectively halving the available bandwidth. Furthermore, coherent detection in optical channels is often plagued by phase noise and high computational requirements.
This paper proposes a streamlined architecture that utilizes Amplitude-Coherent (AC) detection and the Discrete Cosine Transform (DCT) to overcome these hurdles, prioritizing simplicity and robustness.
System Architecture and Methodology
The core of the proposed system is C-OFDM (Cosine-transform OFDM) paired with Amplitude-Coherent (AC) detection. This design deviates from traditional Fourier-based approaches in several key ways:
DCT-based Modulation
Instead of the standard Fast Fourier Transform (FFT), the system employs the Discrete Cosine Transform (DCT). Since the DCT is inherently real-valued, it eliminates the need for the Hermitian symmetry required by optical intensity modulators. This allows every subcarrier to carry data, doubling the spectral efficiency compared to standard DC-biased optical OFDM (DCO-OFDM).
Amplitude-Coherent Detection
The receiver is designed to detect only the amplitude of the received symbols, using unipolar M-ASK modulation. By ignoring the phase information, the system becomes completely immune to phase-related impairments such as phase noise and I/Q imbalance, which are common in high-speed optical links.
Blind Channel Estimation
To avoid the bandwidth overhead associated with pilot symbols, the authors developed a blind channel estimator. This estimator identifies the channel gain ($\alpha$) by sorting received samples and utilizing the highest-energy samples as "virtual pilots." This statistical approach allows for accurate channel state information (CSI) acquisition without sacrificing data subcarriers.
PAPR Reduction via Random Bit-Flipping
High Peak-to-Average Power Ratio (PAPR) is a persistent issue in OFDM systems. The proposed solution is a Random Bit-Flipping Algorithm (RBFA), a greedy method that flips the polarity of data symbols to minimize peaks. Because the AC receiver ignores phase, it does not need to be informed of which bits were flipped, eliminating the need for feedback overhead between the transmitter and receiver.
Performance Evaluation
The proposed design was validated using Monte Carlo simulations with $10^5$ realizations per data point. The system was tested against various channel conditions:
- Flat fading models representing typical VLC environments.
- Gamma-Gamma distribution models simulating atmospheric turbulence in FSO.
- AWGN and Phase Noise models (Tikhonov random variables) to assess robustness.
Key Results and Insights
The study reveals several significant advantages of the AC-DCT architecture:
- Computational Efficiency: The C-OFDM system is remarkably lightweight. The transmitter and receiver require 3x fewer real multiplications and 7x fewer real additions than a conventional Fourier-based system, making it ideal for low-power ubiquitous computing devices.
- Phase Noise Immunity: In comparative simulations, a standard 256-QAM system failed completely even with a low phase noise variance of $1^\circ$. In contrast, the proposed AC-ASK system remained entirely unaffected, demonstrating its suitability for hardware with imperfect oscillators.
- Spectral Gains: By removing the requirement for Hermitian symmetry, the system matches or exceeds the data rates of more complex QAM systems while utilizing simpler hardware.
- Effective PAPR Reduction: The RBFA successfully reduced PAPR by 1 to 4 dB, improving the linear operating range and efficiency of the optical transmitters (e.g., LEDs or laser diodes).
Limitations and Future Work
Despite its advantages, the system has inherent limitations. The use of unipolar ASK results in a smaller minimum Euclidean distance between constellation points compared to QAM, which may lead to higher bit-error rates in very low Signal-to-Noise Ratio (SNR) regimes.
Additionally, the current study assumes perfect time synchronization. Future research must address the impact of synchronization errors on AC detection. There is also an opportunity to refine the blind estimator by developing algorithms that dynamically select the number of "virtual pilots" ($L$) based on real-time channel conditions.
Conclusion
The transition to amplitude-coherent detection represents a paradigm shift for OWC, treating phase as a liability rather than an asset. By combining DCT-based OFDM with blind estimation and phase-insensitive PAPR reduction, this architecture provides a robust, high-efficiency solution for the next generation of optical wireless networks.
Summary updated