Formation-Constrained Cooperative Localization for UAV Swarms in GNSS-Denied Environments

Area of Science:

• Robotics
• Navigation Systems
• Artificial Intelligence

Background:

• Cooperative localization is essential for Unmanned Aerial Vehicle (UAV) swarms in Global Navigation Satellite System (GNSS)-denied environments.
• The backbone-listener scheme optimizes communication and scalability by designating some UAVs as active backbone nodes and others as passive listeners.
• Existing methods may lack accuracy when dealing with complex formation geometries or dynamic environments.

Purpose of the Study:

• To develop a formation-constrained cooperative localization method for UAV swarms in GNSS-denied environments.
• To improve the accuracy, reliability, and stability of UAV swarm localization by leveraging known formation geometry.
• To enhance the adaptability of cooperative localization techniques for practical deployment in diverse formation scenarios.

Main Methods:

• Proposed a formation-constrained greedy node activation (GNA) strategy for backbone node selection, incorporating weighted distance fusion.
• Integrated formation constraints into Chan's algorithm for listener node localization, coupled with angle-of-arrival (AOA) refinement.
• Employed gradient descent-based global optimization to enforce formation constraints across all agent positions in two-dimensional (2D) space.

Main Results:

• The formation-constrained method significantly improved localization accuracy and reliability compared to existing schemes.
• The greedy node activation strategy with weighted distance fusion enabled near-optimal backbone node selection aligned with formation structure.
• Listener localization using Chan's algorithm with AOA refinement and global optimization successfully maintained expected inter-agent distances.

Conclusions:

• The proposed formation-constrained cooperative localization method enhances UAV swarm performance in GNSS-denied conditions.
• The method demonstrates robust adaptability to various formation types, including asymmetric geometries, making it suitable for practical applications.
• Simulation results validate the effectiveness and stability of the approach, highlighting its potential for real-world UAV swarm operations.