Some optimisation-based tightly-coupled multisensor SLAM algorithms
Source: Wu 2018
- Uses nonlinear optimization
- may potentially achieve higher accuracy due to the capability to limit linearization errors through repeated linearization of the inherently nonlinear problem
- [117] Forster 2017 : preintegration theory
- [118] OKVIS: a novel approach
- to tightly integrate visual measurements with IMU
- optimise a joint nonlinear cost function that integrates an IMU error term with the landmark reprojection error in a fully probabilistic manner
- real-time operation: old states are marginalized to maintain a bounded-sized optimization window
- Li et al. [119] proposed tightly coupled, optimization based, monocular visual-inertial state estimation for camera localization in complex environments. This method can run on mobile devices with a lightweight loop closure .
- VIORB a tightly coupled visual-inertial slam system was proposed, following ORBSLAM monocular SLAM