The document discusses Kalman filters, which are algorithms used to estimate the internal state of a system from a series of noisy measurements. Specifically:
- Kalman filters were developed in the 1960s to estimate the state of dynamic systems and filter out noise from sensor measurements in a reliable way.
- They work by using a system's predicted state and measurements from sensors to produce estimates of the true state that are better than either the predictions or measurements alone.
- The algorithm models the dynamic behavior of the system being estimated as well as the measurement noise characteristics, and uses these to produce an optimal estimate.