Kalman filter-based distributed Gaussian process for unknown scalar field estimation in wireless sensor networks

Abstract

This paper proposes an online scalar field estimation algorithm of unknown environments using a distributed Gaussian process framework in wireless sensor networks. While kernel-based Gaussian processes are commonly used for scalar field estimation, their centralized nature makes it difficult to deal with a large number of measurements from sensor networks. To address this, recent advancements approximate kernel functions via E-dimensional nonlinear basis functions, improving scalability. However, this requires many basis functions for accurate approximation, increasing computational and communication complexities. We propose a Kalman filter-based distributed Gaussian process framework, which scales linearly with the number of basis functions and includes a new consensus protocol for efficient data transmission and rapid convergence. Simulation results demonstrate rapid consensus convergence and outstanding estimation accuracy achieved by the proposed algorithm. The scalability and efficiency of the proposed approach are further demonstrated by online dynamic environment estimation using sensor networks.