Given , error covariance, , previous estimate, consensus |
update parameter , and the window size . |
β1. Obtain measurement . |
β2. For each measurement solve L1-norm optimization problem, |
βββreject outliers as given in (3.5) and then obtain the trimmed |
βββmeasurements: . |
β3. Calculate the mode probability . |
βββGiven , |
βββFor |
ββββEvaluate measurement likelihood for . |
ββββEvaluate the Bayesian recursion (3.8)-(3.9). |
βββEnd |
βββDecide the channel mode using threshold testing. |
β4. Compute contribution term of information state and matrix |
βββsuch that |
ββββββββββ, |
βββββββββ. |
β5. Broadcast message to neighbors in . |
β6. Collect messages from neighbors. |
β7. Aggregate the information states and matrices of neighbors |
βββincluding node : : |
βββββββββ. |
β8. Compute the Kalman-Consensus estimate: |
ββββββββββ, |
β. |
βββPrediction stage |
,
|
ββββββββββ . |