We also do the opposite, and we use our IMU to better estimate the position of those features/clues. This is exactly what SLAM does, at its core it is just a Ka

We also do the opposite, and we use our IMU to better estimate the position of those features/clues. This is exactly what SLAM does, at its core it is just a Kalman filter with the twist that each time it finds a reliable visual clue it treats it as a sensor and makes it part of its state, a thing that allow us to tell our position much better and build a map. Some nitpicking here: There are some SLAM systems and flavors that use a Kalman filter as the main estimator under the hood but most are either hybrids or separate the tracking and map optimization (using batch / bundle adjustment methods). Digging a little deeper here I think it's important to mention that a Kalman filter is just a recursive least squares estimator, and is in fact equivalent to an Information Filter (which, IMO, is much easier to understand if you're already familiar with linear regression / LS systems). In fact a Kalman filter is just a fancy mathematical reformulation of an Information Filter that allows you to save quite a bit of computation time in case each observations is low-dimensional compared to your state.