Simple Multiple Target Tracking

This chapter develops a basic multiple target tracker which will be useful in relatively simple multiple target tracking situations. The difficulties in multiple target tracking arise when the number of targets is uncertain and there is ambiguity in deciding which target generated which measurement or whether, in fact, the measurement was generated by a false target. This chapter presents a simplified non-linear Joint Probabilistic Data Association (JPDA) algorithm that assumes measurements arrive one at time. Section 4.2 defines association probability, i.e., the probability that a given target generated a measurement and shows how to calculate it in a Bayesian fashion. Section 4.3 defines hard and soft association in terms of association probabilities. Section 4.4 uses soft association to develop a simplified JPDA tracker and provides an example of its use. It then discusses feature aided tracking and shows how features can aid in associating measurements to targets and improve track estimates. Section 4.5 briefly discusses multiple target tracking methods designed for more complicated problems.