Distributed Kalman Filtering Based on the Non-Repeated Diffusion Strategy

. 2020 Dec 3;20(23):6923. doi: 10.3390/s20236923

Algorithm 5: The non-repeated diffusion strategy.
For the node $i \in V$ and node $i_{1} \in N_{i}$ , Initialize information of node i at time t with $I_{i, t}^{loc}$ ; Initialize with: $I_{i_{1} \to i, 0} = 0$ ; Diffusion update:
$I_{i \to i_{1}, t} = \sum_{i_{2} \in N_{i} - i_{1}} I_{i_{2} \to i, t - 1} + I_{i, t}^{loc}$ ;	(8)
Communication and fusion information: Send $I_{i_{1} \to i, t}$ to adjacent node i,
$I_{i, t} = \sum_{i_{1} \in N_{i}} I_{i_{1} \to i, t} + I_{i, t}^{loc} .$	(9)