The initial release provides the library foundation which contains a filter-based visual-inertial localization solution. This can be used in a wide range of scenarios and the type-based index system allows for others to easily add new features and develop on top of this system. Here is a list of future directions, although nothing is set in stone, that we are interested in taking:
- Creation of a secondary graph-based thread that loosely introduces loop closures (akin to the second thread of VINS-Mono and others) which should allow for drift free long-term localization.
- Large scale offline batch graph optimization which leverages the trajectory of the ov_msckf as its initial guess and then optimizes both the map and trajectory (see ov_maplab project!).
- Incorporate our prior work in preintegration [Eckenhoff2019IJRR] into the same framework structure to allow for easy extensibility. Focus on sparsification and marginalization to allow for realtime computation.
- Support for arbitrary Schmidt'ing of state elements allowing for modeling of their prior uncertainties but without optimizing their values online.