Input: The raw data database at the th time stamp |
Output: The aggregation secure statistics |
Find out the optimal number of sampling points |
Find out sets of sampling sensors at the current time stamp |
Obtain the grouping strategy via smart grouping |
Allocate the budget for all sampling sensors |
Add Laplacian noise to group with allocated budget at the perturbation mechanism |
Report the aggregated secure statistics that is filtered by Kalman filtering. |
Update the sampling interval by adaptive sampling |