Distributed networks estimate states and unknown dynamics with limited data
COIN-GP: Cooperative Online Learning in Networked Distributed Systems with Partial Measurements via Gaussian Process Regression
Machine LearningMultiagent Systems
Summary
When many sensors work together to understand a system, they often can't see everything they need to know. This paper presents a new way for these sensors to learn about the system and its hidden behaviors even when they have incomplete information. The authors designed a method that uses a form of smart prediction called Gaussian Process regression and a clever way to gather data to improve estimates. They also provide mathematical guarantees on how accurate these predictions can be.
What this means in practice
- •For industrial sensor network engineers: Enable sensor teams to jointly estimate full system states despite incomplete sensor data using an online cooperative approach.
- •For smart building system integrators: Improve building monitoring by estimating partially observed environmental or system variables through distributed sensor collaboration.
Authors
Zewen Yang, Xiaobing Dai, Zhenxiao Yin, Hang Zhao, Zhijun Li, C. C. Chan
Abstract
In this paper, we tackle the problem of jointly estimating the system states and partially unknown dynamics within distributed sensor-equipped networks, particularly in scenarios where only partial state observations are available. To address this issue, we propose an observer-based dynamic cooperative learning framework incorporating online distributed Gaussian Process (GP) regression, which enables accurate estimation despite incomplete in measurements and deficient GP models. In addition, a novel data collection strategy is introduced, with theoretical conditions ensuring feasible data acquisition. Moreover, we also derive an error upper bound encompassing state estimation and model estimation, leveraging the deterministic error bounds of GPs. Empirical simulations demonstrate the superiority of our approach compared to existing distributed GP-based methods.