State feedback, state estimation
Feedback is the most important element in operating systems; it is used by all creatures and technical control systems. In essence, it compares the measured output, performance characteristics or states of the system with the desired characteristics, relations and influences the states of the system accordingly. The control signal is built as a weighted sum of the measured state variables. Feedback reverses the relation between output and input; in computational terms, the output signals are inverted according to the inverse of the system model. The feedback loop can include elements that accelerate or slow down operation, monitor average values, or sensitively react to major changes. It is a separate task to filter out (isolate from normal operation) the effect of the disturbances which influence system operation. This task is usually fulfilled with the Kalman filter and its advanced versions. Basically, the Kalman filter monitors the states of the process and separates normal and disturbing changes on a statistical probability basis benefitting from the fact that normal operation and disturbances have different statistical characteristics.