Decentralized State Estimation

Decentralized state estimation focuses on developing methods for multiple agents or sensors to collaboratively estimate a shared state without relying on a central processing unit, improving robustness and scalability. Current research emphasizes algorithms like Extended Kalman Filters and Moving Horizon Estimation, often combined with techniques like pseudomeasurements and preintegration to efficiently share information and handle constraints. This approach is crucial for applications ranging from multi-robot systems and swarm robotics to indoor localization and tracking, offering advantages in terms of resilience, reduced communication overhead, and enhanced privacy.