Data Association

Abstract

In tracking applications, following the signal detection process that yields measurements, there is a procedure that selects the measurement(s) to be incorporated into the state estimator – this is called data association (DA) . In multitarget-multisensor tracking systems, there are generally three classes of data association: specifically, measurement-to-track association (M2TA) , track-to-track association (T2TA) , and measurement-to-measurement association (M2MA) . M2TA is the process of associating each measurement from a list (originating from one or more sensors) to a new or existing track. T2TA is the process of associating multiple existing tracks (from multiple sensors or from different periods in time), generally with the intent of fusing them afterward. M2MA is the process of associating measurements from different sensors in order to form “composite measurements” and/or do track initialization. The processes of M2TA and T2TA will be discussed in more detail here, while details on M2MA can be found in Bar-Shalom et al. (2011).