Multi-Dimensional Time Series Research Data Normalization

Develop a SQL solution for normalizing and aggregating time-series experimental data from multiple scientific instruments with varying sampling rates and timestamp formats. Create a robust PostgreSQL function that can handle: 1) Different time precision levels, 2) Missing data interpolation, 3) Outlier detection using statistical methods, and 4) Automatic schema adaptation for new data sources. The solution must support millisecond-level precision and handle datasets with potential measurement errors.