Hermite accelerated Monte-Carlo for evaluating multitarget tracking filters

A new method is proposed for the statistical evaluation of multitarget tracking filter performance. The method is based on accelerated Monte-Carlo simulation, with numerical convergence hastened by the use of orthonormal function-space expansions. In particular, use is made of the Hermite polynomial functions. This new method is likened to numerical quadrature except for the type of smoothing performed. This new technique does global smoothing and for optimum efficiency is adapted to the general mathematical properties of the particular class of problems being considered.