ABSTRACT
Within the overall context of multi-sensor data fusion, Kalman filters provide a classic sequential estimation approach for fusion of kinematic and attribute parameters to characterize the location, velocity and attributes of individual entities (e.g., targets, platforms, events, or activities). This chapter (based on material originally presented in Chapter 4 of Hall [1992]) provides an introduction to estimation, sequential processing, and Kalman filters. The problem of fusing multi-sensor parametric data (from one or more sensors) to yield an improved estimate of the state of the entity is a classic problem. Examples of estimation problems include
Using positional data such as line-of-bearing (angles), range or range-rate observations to determine the location of a stationary entity (e.g., determining the location and velocity of a ground-based target using a distributed network of ground sensors)
Combining positional data from multiple sensors to determine the position and velocity of a moving object as a function of time (the tracking problem)
Estimating attributes of an entity, such as size or shape, based on observational data
Estimating the parameters of a model (e.g., the coefficients of a polynomial), which represents or describes observational data
