LSLOCK: A Method to Estimate State Space Model by Spatiotemporal Continuity

Model estimation from spatio-temporal data is important topic since it helps us to extract useful information from big data in recent years. In this paper, we introduce an estimation algorithm of the linear Gaussian state space model with focusing on the real-time property. The proposed algorithm uses two key ideas, localization and spatial uniformity, to reduce the number of the parameters. Thanks to this, we obtain stable method to estimate the parameters regarding state transition and states. In addition, the proposed algorithm is quicker and more accurate than existing methods, therefore, it suffices the requirement of the rapid response for the alternation of the fields.