HF RL Explorer

Implement Kalman smoother using Rauch-Tung-Striebel algorithm

Implement Kalman smoother using Rauch-Tung-Striebel algorithm: a task in LegoFlow-SWE (Harbor dataset). When a Kalman filter processes measurements forward, the final state estimate only uses information up to the last measurement. To obtain state estimates at each measurement time that…

The task

When a Kalman filter processes measurements forward, the final state estimate only uses information up to the last measurement. To obtain state estimates at each measurement time that incorporate **all** future measurements, a backward smoothing pass is needed – the Rauch-Tung-Striebel (RTS) algorithm.

Part of Lego-X/LegoFlow-SWE.