Bounds on track purity for multiple target tracking

Theoretical bounds are derived on the probability of correct data association in a dense multiple-target environment with clutter, false alarms, nonunity probability of detection, etc. The theory allows the best possible performance to be estimated without implementing the optimal algorithm. The bounds can be used to study the fundamental limitations of tracking performance, rather than pursuing an endless quest for better algorithms. For some applications, however, the use of the theory could show that there is substantial room for algorithmic improvement