Noise Mitigation Methods for Digital Visible Light Communication

Visible Light Communication (VLC) using Light Emitting Diodes (LEDs) has gained attention due to its low power consumption, long lifetime, and fast response. However, VLC suffers from optical noise generated by ambient light sources such as fluorescent lamps, which leads to waveform distortion and increased bit error rates (BER). In this paper, we propose two noise reduction methods for Digital Visible Light Communication (DVLC) systems. The first method exploits the periodic nature of interference caused by AC-powered-line illumination and reduces interference by subtracting sampled noise waveforms from the received signal. Second, inspired by Active Noise Control (ANC) techniques, an additional photodiode is introduced for noise reception, and subtraction circuits are employed to attenuate noise in real time. Experimental results show that both methods improve BER performance compared with conventional receivers, with the ANC-inspired approach achieving superior performance under all tested conditions.

💡 Research Summary

The paper addresses a critical obstacle to the practical deployment of digital visible light communication (DVLC) systems that use LEDs: optical noise generated by ambient lighting such as fluorescent lamps. This noise, which originates from the rectification of AC power and manifests as a periodic waveform at twice the mains frequency (100 Hz or 120 Hz depending on the region) together with transient spikes, distorts the received voltage levels and leads to erroneous HIGH/LOW decisions in the receiver’s comparator, thereby increasing the bit‑error rate (BER).

After a concise review of VLC fundamentals—including the visible spectrum (380–780 nm), the fast switching capability of LEDs, and digital modulation schemes (OOK, PPM, I‑PPM)—the authors categorize noise into two groups. Periodic noise is directly linked to the mains frequency and is highly predictable, while non‑periodic noise includes rapid illumination changes, multi‑source interference, and other stochastic disturbances. Conventional low‑pass, high‑pass, or band‑pass filters are shown to be insufficient when the noise spectrum overlaps the signal band, especially at higher data rates where multiple symbols fall within a single noise cycle.

To mitigate these effects, the authors propose two complementary techniques.

1. Periodic‑Interference Subtraction: During idle periods, a single cycle of the interference waveform is sampled with an analog‑to‑digital converter (ADC) and stored as a reference. For subsequent data reception, this stored waveform is subtracted from the incoming noisy signal before it is converted back to analog (via a DAC) and fed to the microcontroller for demodulation. This method exploits the repeatability of mains‑derived interference and requires only modest additional circuitry.
2. ANC‑Inspired Real‑Time Subtraction: An auxiliary photodiode is positioned to capture only the ambient light component, acting as a “noise reference.” The reference signal is fed into a subtraction circuit (either an analog differential amplifier or a digital subtractor) that removes the noise component from the main photodiode’s output in real time. This approach is analogous to active noise control in acoustics and is capable of attenuating both periodic and non‑periodic disturbances.

Experimental validation was performed under a 60 Hz power‑line environment (typical for western Japan) at two data rates: 10 kbps and 2.5 kbps. The performance metric was the relationship between energy‑per‑bit‑to‑noise ratio (E_b/N_0) and BER. Both proposed methods outperformed a baseline receiver that employed only conventional filtering. The ANC‑inspired technique consistently achieved the lowest BER across all tested conditions, demonstrating superior robustness to both types of noise. Quantitatively, the methods yielded approximately 1–2 dB improvement in required E_b/N_0 for a given BER target.

The authors acknowledge several limitations. The experiments were confined to a controlled indoor setting; outdoor sunlight, rapidly varying illumination, and multi‑user scenarios were not examined. The additional photodiode and subtraction hardware increase system cost, board space, and power consumption, and precise optical alignment and timing synchronization between the two sensors are required. Moreover, the paper does not provide a detailed analysis of processing latency or energy overhead, which are critical for battery‑powered or mobile DVLC devices.

In conclusion, the study contributes two practical noise‑mitigation strategies for DVLC. The periodic‑subtraction method offers a low‑cost solution when the interference is highly repeatable, while the ANC‑inspired real‑time subtraction delivers higher performance at the expense of added complexity. Future work should explore adaptive filtering algorithms (e.g., LMS, RLS), machine‑learning‑based noise modeling, and low‑power ASIC implementations to integrate these techniques into compact, energy‑efficient VLC transceivers suitable for real‑world deployments.