Point process formalism for multiple target tracking

This paper discusses the point process formalism of multiple target tracking problems. Finite point processes are defined as random elements in the spaces of finite sequences with their orders ignored, rather than as random integer-valued measures. Two recently developed algorithms, Joint Multitarget Probability (JMP) algorithm and Unified Data Fusion (UDF) algorithm will be described as algorithms based on the point process formalism. Two more established algorithms, Multiple Hypothesis Tracking (MHT) and Probabilistic Data Association (PDA) algorithms, will also be examined as algorithms using the point process concept. The point process formalism will then be compared with the random set formalism developed by Ron Mahler (2001).