Data association algorithm for radar and infrared sensor based on information divergence

The cost function of the current multi-target data association algorithms based on the S-D assignment for heterogeneous sensors is computed by using maximum likelihood estimation of the target position without taking the estimation errors into account. To overcome the problem above, a new data association algorithm was proposed based on the information divergence,which first used the unscented transform to get the statistic information of the pseudo measurements,then the differentia between the probability density function of pseudo measurements and the most posterior probability density function worked as the association cost. Finally simulation experiments were made to validate the proposed algorithm. The results show that the proposed algorithm can achieve better performance and its association cost reflects the association probability more accurately.