Kalman filtering (or linear quadratic estimation) is an algorithm to estimate statistics of interest via multiple, parallel noisy time series measurements. It yields estimates that are more accurate than those based on a single measurement by estimating a joint probability distribution over the variables for each time-step.
An example is the aggregation of measurements of Inertial Measurement Unit (IMU) data, odometer and GPS into a more accurate estimate of distance travelled for a car than any single one of those information sources.
Resources A very complete multipart tutorial on the Kalman filter can be found at kalmanfilter.net. There is a mathematically rigorous tutorial on the Kalman filter from Tony Lacey hosted by MIT. How a Kalman filter works, in pictures gives a more fun and visually appealing intro.