ALTERNATIVES TO MONTE-CARLO SIMULATION EVALUATIONS OF 2 MULTISENSOR FUSION ALGORITHMS

Parallel and sequential multisensor extensions of the probabilistic da ta association filter for multitarget tracking in clutter are presente d, and a non-simulation technique is developed and used to compare tra cking performance of the two algorithms in case of a single target. Wh ile requiring only a fraction of the time for Monte-Carlo simulation e valuation, the non-simulation technique is shown to accurately predict the average superior performance of the sequential implementation in terms of RMS position error and track lifetime which has been observed in simulations. (C) 1998 Elsevier Science Ltd. All rights reserved.