This chapter was written by Martin P. Dana, Raytheon Systems, Retired

State estimation as it relates to object tracking is an important element of Level 1 fusion. While many facets of this topic were introduced in Chapter 3, here we delve further into several areas that are critical to the implementation of modern multi-sensor tracking systems that incorporate data fusion as part of the state-estimation process. These include discussions of the general design approaches and implementations for several of the fundamental elements of a radar tracking logic. Signal and data processing found in a radar tracker may need to account for the unique characteristics of measurement data, state estimates (tracks), or both depending on the output of the radar subsystems. The design must also incorporate measures of quality for tracking and tracker performance, and the ability to measure and account for sensor registration errors that exist in a multi-sensor tracking system. Other issues addressed in the chapter include the transformation of radar measurements from a local coordinate system into a system-level or master coordinate system, standard and extended Kalman filters, track initiation in clutter, state estimation using interacting multiple models, and the constraints often placed upon architectures that employ multiple radars for state estimation.