Optical communication-based identification for multi-UAV systems: theory and practice

Basic Information

• Paper ID: 2302.04770
• Title: Optical communication-based identification for multi-UAV systems: theory and practice
• Authors: Daniel Bonilla Licea, Viktor Walter, Mounir Ghogho, Martin Saska
• Classification: eess.SP (Electrical Engineering and Systems Science - Signal Processing)
• Published Journal: Autonomous Robots (2025)
• Paper Link: https://arxiv.org/abs/2302.04770

Abstract

Mutual localization and identification in multi-UAV systems are essential features for stability and navigation. Camera-based communication technology, namely Optical Camera Communication (OCC), provides a valuable solution to this complex task. In this system, UAVs are equipped with LED beacons and cameras to achieve mutual identification through specific flickering sequences. The system is immune to RF electromagnetic interference, can operate in GNSS-denied environments, and has low implementation costs. This paper conducts an in-depth investigation of the system's capacity and limitations, demonstrates how to construct LED flickering sequences to enhance system performance, and validates the theoretical analysis through experiments.

Research Background and Motivation

Problem Definition

Multi-UAV systems need to address the core challenges of mutual localization and identification, which are critical for:

1. Cooperative Flight: Enabling tight collaborative flight and mutual collision avoidance
2. High-Level Planning: Identifying nearby team members is essential for advanced task planning
3. Environmental Adaptability: Maintaining functionality in GNSS-denied and RF-interference environments

Limitations of Existing Methods

Traditional approaches suffer from the following issues:

1. RF Methods: RTK-GNSS, UWB ranging, etc., are susceptible to electromagnetic interference
2. Passive Vision Systems: Sensitive to ambient lighting and fail in low-light environments
3. Frequency Separation Methods: Using different frequencies to distinguish LED signals is highly inefficient, producing only L/2 distinct sequences

Research Motivation

This paper proposes an optical communication-based identification scheme based on the UVDAR (UltraViolet Direction And Ranging) system, featuring:

• Immunity to RF electromagnetic interference
• Operation in GNSS-denied environments
• Low implementation cost (utilizing existing camera equipment)
• Active optical systems overcoming lighting challenges

Core Contributions

1. LED Flickering Sequence Generation Method: Develops a theoretical framework for designing flickering sequence sets for UAV LED arrays, optimizing sequences to distinguish as many sequences as possible in minimal time
2. UVDAR Theoretical Analysis: Derives analytical expressions for the false detection probability of flickering sequences and determines the relationship between sequence length and the number of detectable sequences
3. Experimental Validation: Implements a prototype visual mutual identification system and conducts outdoor testing for verification

Methodology Details

Task Definition

Input: LED flickering signals from each UAV in the multi-UAV system Output: Identification of the UAV ID corresponding to the signal Constraints: Minimize identification time, maximize the number of identifiable UAVs, ensure system robustness

System Architecture

1. Clock Signal Model

The k-th falling edge time of the j-th UAV is modeled as:

t_{j,k} = T_j + n_{j,k} + t_{j,k-1}

where T_j is the true clock period and n_{j,k} represents frequency instability.

2. Optical Transmitter

Contains M parallel branches (M=4 in this paper), each branch includes:

• Binary Stream Generator: Generates repeated binary sequences of length L
• Encoder/Modulator: Uses NRZ or Manchester encoding with OOK modulation
• Analog Front-End: Converts electrical signals to optical signals v_{j,m}(t) = P·u_{j,m}(t)

3. Optical Receiver

• Camera: Grayscale camera equipped with UV filter, exposure time τ_e
• Image Processing: Detects bright spots, tracks motion, extracts optical signal time series
• Classifier: Determines identification number through correlation with sequences in dictionary D

4. Optical Channel Model

x_ℓ[k] = h_ℓ(t_k)∫_{t_{ℓ,k}}^{t_{ℓ,k}+τ_e} v_j(t)dt + n_ℓ[k]

Binary Sequence Design

Sequence Requirements

1. Minimum Average Power: ||b_n||_0 ≥ b̄L, ensuring LED detection and tracking
2. Limit Consecutive On-Time: Maximum N_1 consecutive '1' bits to distinguish reflected light
3. Limit Consecutive Off-Time: Maximum N_0 consecutive '0' bits to prevent tracking failure
4. Cyclic Equivalence: Handle time desynchronization between transmitter and receiver
5. Hamming Distance: Increase cyclic Hamming distance to improve noise robustness

Sequence Generation Algorithm

Algorithm 1 (NRZ Encoding):

1. PowerTest(S_L, b̄)           # Power test
2. CircularityTest(A_L)        # Circularity test  
3. OnesTest(B_L, N_1)          # Consecutive ones test
4. ZerosTest(C_L, N_0)         # Consecutive zeros test
5. HammingTest(D_L, H_m)       # Hamming distance test

Technical Innovations

1. Optimized Sequence Design: Significantly increases sequence capacity compared to simple frequency separation
2. Clock Mismatch Analysis: First systematic analysis of clock signal effects on optical identification systems
3. Practical Considerations: Balances LED power, tracking performance, and interference resistance

Experimental Setup

Hardware Platform

• UAVs: Based on Holybro X500 platform, arm length 0.245m
• UVDAR System: 3 UV cameras + 4 pairs of UV LEDs
• LED Specifications: 395nm wavelength, 600mW driving power, 276mW radiant power
• Camera Configuration: 60fps, fisheye lens, 180° horizontal × 120° vertical field of view

Experimental Scenarios

1. Indoor Static Testing: 5m distance, analyzing clock mismatch effects
2. Outdoor Dynamic Testing: 3-UAV formation flight, 235-second recording

Evaluation Metrics

• Identification Time T_d: Time from signal appearance to successful identification
• Classification Error Probability p_: Probability of misidentification
• Detection Success Rate: Probability of correct identification

Experimental Results

Main Results

1. Hamming Distance Impact Analysis

Comparison of two configurations (J=11 UAVs, 2 sequences each):

• Case A: H_m=1, L=8
• Case B: H_m=3, L=13

2. Clock Mismatch Impact

With clock deviation δ=0.01:

• Identification time increases slightly
• At low SNR, Hamming distance provides significant improvement
• At high SNR, clock mismatch becomes the primary limiting factor

3. System Capacity Analysis

For clock with T/σ_T = 10^4:

• Maximum sequence length L_ decreases as the number of UAVs J increases
• When J≥29, normal operation of all optical links cannot be guaranteed
• Hamming distance H_m=3 requires longer sequences than H_m=1

Outdoor Flight Experiments

• UAV-1 Detection Success Rate: 93.11%
• UAV-2 Detection Success Rate: 63.27%
• Most errors persist for a single sampling period (66.33% and 47.02%)
• System performs well in actual dynamic environments

Clock Characteristic Analysis

Experiments reveal that inter-frame camera intervals follow a Laplace distribution with non-flat power spectrum, confirming the accuracy of the clock instability model.

Related Work

Localization Technology Classification

1. RF Methods: RTK-GNSS, UWB ranging, motion capture systems
2. Passive Vision: Specific marker pattern recognition
3. Active Vision: Infrared LED frequency separation, UV LED systems

Optical Camera Communication

OCC technology is primarily applied to vehicle communication. This paper is the first to systematically apply it to UAV identification, filling a gap in the field.

Conclusions and Discussion

Main Conclusions

1. Sequence Design is Critical: Optimized binary sequences significantly enhance system performance
2. Clock Limitations: Clock signal mismatch is the primary limiting factor for system capacity
3. Practical Validation: The system demonstrates good performance in real environments

Limitations

1. Effective Range: Approximately 15 meters effective identification range with current configuration
2. Environmental Dependence: Strong light sources such as sunlight may affect performance
3. Clock Requirements: Relatively high clock precision requirements

Security Considerations

The system is vulnerable to optical attacks: large amounts of UV light sources may saturate the receiver. Mitigation strategies include increasing flight altitude or employing resource management strategies.

In-Depth Evaluation

Strengths

1. Theoretical Completeness: Complete theoretical framework from sequence design to system analysis
2. Practical Value: Addresses real-world needs in GNSS-denied environments
3. Comprehensive Experiments: Multi-scenario validation both indoors and outdoors
4. Strong Innovation: First systematic study of optical communication applications in UAV identification

Weaknesses

1. Scalability Limitations: Applicability to large-scale UAV swarms requires further verification
2. Environmental Robustness: Performance under complex lighting conditions needs further investigation
3. Computational Complexity: Real-time processing computational overhead not thoroughly analyzed

Impact

1. Academic Contribution: Establishes theoretical foundation for optical communication applications in robotic systems
2. Practical Value: Provides solutions for military, rescue, and other GNSS-denied scenarios
3. Technology Promotion: Extensible to other multi-robot systems

Applicable Scenarios

• Military reconnaissance and combat environments
• Indoor warehousing and logistics
• Search and rescue missions
• Urban canyon environments with GPS signal interference

References

The paper cites 28 related references covering key areas including multi-UAV systems, optical communication, and visual localization, providing a solid theoretical foundation for the research.

Overall Assessment: This is a high-quality systematic research paper with in-depth theoretical analysis and comprehensive experimental validation, making significant contributions to the field of optical identification for multi-UAV systems. This work not only addresses practical needs but also provides important reference for future research.