High-Sensitivity Free Space Optical Communications Using Low SWaP Hardware

Overview

This work demonstrates a practical free-space optical (FSO) communication link leveraging highly integrated, low Size, Weight, and Power (SWaP) hardware. The system combines a Gallium Nitride (GaN) micro-LED transmitter, controlled by a CMOS driver, with a receiver based on a CMOS-integrated array of Single-Photon Avalanche Diodes (SPADs). Using a simple Return-to-Zero On-Off Keying (RZ-OOK) modulation scheme, the link achieves a data rate of 100 Mb/s with a receiver sensitivity of -55.2 dBm (corresponding to ~7.5 detected photons per bit) while consuming less than 5.5 W total power. This represents a significant step towards deployable, high-performance optical communication systems for constrained environments.

100 Mb/s

Demonstrated Data Rate

-55.2 dBm

Receiver Sensitivity @ 100 Mb/s

< 5.5 W

Total System Power

7.5 photons/bit

Detection Efficiency

1. Introduction

Free-space optical communication offers high bandwidth potential but often relies on bulky, power-hungry equipment like external modulator-driven lasers and cryogenic receivers. The drive for applications in small satellites (CubeSats), unmanned aerial vehicles (UAVs), and portable ground terminals necessitates a paradigm shift towards Low-SWaP hardware. This paper addresses this need by exploiting two key CMOS-compatible technologies: high-bandwidth micro-LEDs for transmission and SPAD arrays for ultra-sensitive reception. The integration of both elements into compact, digitally-interfaced systems is the core innovation, moving beyond benchtop demonstrations to practical implementations.

2. Methods & System Architecture

The communication system is built from two integrated subsystems: a transmitter and a receiver, both designed for minimal SWaP.

2.1 Transmitter: CMOS-Controlled Micro-LED

The source is a GaN-based micro-LED, bump-bonded to a CMOS control chip. This integration allows direct digital control of the light emission with high spatial and temporal precision, eliminating the need for separate Digital-to-Analog Converters (DACs) and arbitrary waveform generators. Micro-LEDs offer high modulation bandwidths (capable of Gb/s rates), making them suitable for high-speed communication.

2.2 Receiver: SPAD Array

The receiver core is a CMOS-fabricated array of Single-Photon Avalanche Diodes. A SPAD operates in Geiger mode, producing a detectable electrical pulse upon absorption of a single photon, followed by a dead time. Arraying SPADs and combining their outputs mitigates dead time limitations and enables high dynamic range. The CMOS integration allows for significant on-chip signal processing (e.g., quenching, counting), reducing backend complexity.

2.3 Modulation Scheme: RZ-OOK

The chosen modulation is Return-to-Zero On-Off Keying. While requiring more bandwidth than Non-Return-to-Zero (NRZ), RZ-OOK reduces Inter-Symbol Interference (ISI) in SPAD-based systems caused by dead time and photon arrival statistics. The signal is decoded using a simple threshold detector. The photon detection process is Poissonian. The probability of detecting k photons in a bit period with an average arrival rate of $\lambda$ photons/bit is given by: $$P(k) = \frac{e^{-\lambda} \lambda^k}{k!}$$ The Bit Error Ratio (BER) is fundamentally limited by this statistics towards the Standard Quantum Limit (SQL).

3. Experimental Results & Performance

3.1 Sensitivity & Data Rate

The primary results are summarized in the performance metrics. The link demonstrated two key operating points:

50 Mb/s: Achieved a sensitivity of -60.5 dBm.
100 Mb/s: Achieved a sensitivity of -55.2 dBm, corresponding to approximately 7.5 detected photons per bit.

This 100 Mb/s sensitivity is reported to be 18.5 dB away from the Standard Quantum Limit (SQL) of -70.1 dBm for 635 nm light, indicating room for further improvement through advanced coding and detection algorithms.

3.2 Power Consumption & SWaP Metrics

A critical achievement is the total system power consumption of less than 5.5 Watts for the unoptimized prototype. This low power, combined with the inherent compactness of CMOS and flip-chip bonded devices, validates the low-SWaP premise. The system foregoes power-hungry components like thermo-electric coolers (common in APDs) or cryogenic systems (for superconducting detectors).

3.3 Bit Error Ratio Analysis

BER curves were measured as a function of received optical power. The curves show the characteristic steep slope of photon-counting receivers. Performance degradation at higher data rates is attributed to increased impact of SPAD dead time and ISI. The use of RZ modulation provided a clear BER advantage over NRZ in this context, as predicted.

Chart Description (Implied): A graph plotting BER (log scale) versus Received Optical Power (dBm). Two curves are shown for 50 Mb/s and 100 Mb/s. The 50 Mb/s curve reaches a BER of 1e-3 at a lower power (more sensitive) than the 100 Mb/s curve. Both curves show a sharp "waterfall" region. Dotted lines may indicate the theoretical SQL limit.

4. Technical Analysis & Core Insights

Core Insight: This paper isn't about breaking pure sensitivity records; it's a masterclass in pragmatic systems engineering. The real breakthrough is proving that near-quantum-limited sensitivity (-55.2 dBm at 100 Mb/s) can be extracted from a dirt-simple, digitally-native, and miserably low-power (<5.5W) box. While others chase dB closer to the SQL with liquid helium and complex DSP, Griffiths et al. ask: "What good is a -70 dBm link if it needs a truck to carry it?" Their answer integrates a micro-LED and SPAD array directly onto CMOS, turning what was a lab curiosity into a deployable asset for SWaP-constrained platforms like CubeSats and drones.

Logical Flow: The argument is elegantly linear. 1) High-sensitivity FSO exists but relies on bulky, high-power hardware (problem statement). 2) Two CMOS-compatible technologies—micro-LEDs (fast, integrable transmitters) and SPAD arrays (single-photon sensitive, integrable receivers)—are identified as solutions. 3) Integrate them into a minimal system using the simplest possible modulation (RZ-OOK) to avoid complex, power-hungry coding. 4) Measure: the data shows high sensitivity and low power simultaneously. The logic proves integration + simplicity = practical high performance.

Strengths & Flaws: The strength is undeniable: demonstrated system-level SWaP efficiency that is rarely measured, let alone achieved, in academic photonics papers. The choice of RZ-OOK is clever for mitigating SPAD dead-time issues. However, the flaw is in the trade-off made for that simplicity. A 100 Mb/s rate is modest, and the 18.5 dB gap to the SQL is significant. As noted in seminal SPAD communication work like D. Chitnis and S. Collins, "A SPAD-based photon detecting system for optical communications," JLT 2014, advanced modulation (e.g., PPM) and forward error correction could close much of that gap. The paper acknowledges this but leaves it for future work, slightly weakening its claim of optimality.

Actionable Insights: For industry, this is a blueprint: stop over-designing. Start with deeply integrated photonic-electronic cores (CMOS is your friend) and only add complexity (modulation, coding) if the simple solution fails. The <5.5W power budget is the number to beat for next-gen product managers. For researchers, the path is clear. The next paper must bridge the sensitivity gap using on-chip coding and processing. Can low-power CMOS logic implement near-capacity codes like LDPC to recover those 18 dB? That's the billion-dollar question for making this technology dominant in 6G backhaul or satellite constellations, moving beyond niche applications.

5. Analysis Framework & Case Example

Framework: SWaP-Constrained System Design Trade-off Matrix

This case exemplifies a structured trade-off analysis for embedded photonic systems. The framework prioritizes constraints and makes deliberate sacrifices.

Primary Constraint Identification: SWaP is paramount. This immediately rules out high-power lasers, external modulators, cryogenics, and bulky discrete optics.
Technology Selection (The "What"): Map required functions (high-speed emission, single-photon detection) to the most SWaP-efficient, integrable technologies: Micro-LEDs and CMOS SPADs.
Complexity Minimization (The "How"): Choose the simplest algorithm/modulation that meets the core performance spec. Here, maximum sensitivity at a target data rate (100 Mb/s) is the goal, not maximum spectral efficiency. Hence, complex m-QAM is rejected in favor of simple RZ-OOK.
Integration Point Definition: Define the boundary where custom hardware must take over from software to save power. Here, the photon counting and basic thresholding are pushed into the CMOS SPAD array's dedicated circuitry.
Metric Validation: Measure the full system against all primary constraints (Sensitivity: -55.2 dBm, Power: <5.5W, Data Rate: 100 Mb/s), not just the optimal performance of a sub-component.

Case Application: The authors applied this framework perfectly. They sacrificed spectral efficiency and ultimate sensitivity (accepting the 18.5 dB gap to SQL) to win on the primary constraints of power and integrability. A contrasting failed approach would be taking a high-sensitivity superconducting nanowire single-photon detector (SNSPD) and trying to miniaturize its cryocooler—a fight against physics. This paper's success lies in choosing battles it could win with CMOS.

6. Future Applications & Development Directions

The demonstrated technology opens doors to several critical application domains and suggests clear paths for evolution.

CubeSat & Small Satellite Constellations: The ultimate low-SWaP environment. Such links can enable high-speed inter-satellite links (ISL) for mega-constellations, reducing reliance on RF with its spectrum limitations. Companies like SpaceX (Starlink) and Planet Labs are potential end-users.
Unmanned Aerial Vehicle (UAV) Swarms: Secure, high-bandwidth communication between drones for coordinated missions without detectable RF emissions.
Last-Mile Ground Communication: In disaster recovery or military operations, rapid deployment of high-bandwidth links between temporary nodes.
Future Development Directions:
1. On-Chip Coding & DSP: Integrating advanced forward error correction (e.g., LDPC, Polar codes) and detection algorithms directly into the receiver CMOS to close the sensitivity gap to the SQL without increasing power or size substantially.
2. Wavelength Scaling: Moving from 635 nm to telecom wavelengths (1550 nm) for better atmospheric transmission and eye safety, using materials like InGaAs/InP SPADs (though integration with CMOS is more challenging).
3. Beam Steering & Tracking Integration: Incorporating Micro-Electro-Mechanical Systems (MEMS) mirrors or liquid crystal-based beam steerers on the same package for robust alignment in dynamic FSO links, a critical step for mobile platforms.
4. Network Prototyping: Moving from point-to-point links to demonstrating small, ad-hoc networks of these low-SWaP nodes, addressing protocols and network management.

7. References

Griffiths, A. D., Herrnsdorf, J., Almer, O., Henderson, R. K., Strain, M. J., & Dawson, M. D. (2019). High-sensitivity free space optical communications using low size, weight and power hardware. arXiv preprint arXiv:1902.00495.
Chitnis, D., & Collins, S. (2014). A SPAD-based photon detecting system for optical communications. Journal of Lightwave Technology, 32(10), 2028-2034.
Kahn, J. M., & Barry, J. R. (1997). Wireless infrared communications. Proceedings of the IEEE, 85(2), 265-298.
DARPA. (n.d.). Direct On-chip Digital Optical Synthesizer (DODOS) Program. Retrieved from https://www.darpa.mil/program/direct-on-chip-digital-optical-synthesizer
NASA. (2020). Optical Communications and Sensor Demonstration (OCSD). Retrieved from https://www.nasa.gov/smallsat-institute/sst-soa/communications
Richardson, D. J., Fini, J. M., & Nelson, L. E. (2013). Space-division multiplexing in optical fibres. Nature Photonics, 7(5), 354-362.