Optimization-Based Risk-Averse State Estimation for GNSS outlier accommodation in the urban environment
Current targeted signal: GPS: L1 GLO: GAL: BDS: B1I (Before customizing the satellite position calculation, please read BDS ICD; different signal channels may have different computations.)
The Risk-Averse Optimization problem is an multi-convex problem. The convex reformulation of the binary case is evaluated for achieving a globally optimal solution and used as a base comparison. The convex form can be solved by CVX but its runtime is not suitable for real-time applications. The Block Coordinate Method for both binary and non-binary cases results in a locally optimal solution with real-time feasibility and achieves comparable results with the globally optimal solution.
This repo is for the experiment of the PhD dissertation and DGNSS Risk-Aver paper. More information will be added later.