Track fitting with ambiguities and noise: a study of elastic tracking and nonlinear filters

We present a study of track fitting in the presence of ambiguous measuremen ts and noise. We consider four methods for solving the resulting assignment problem, two elastic arm algorithms and two nonlinear filters, including a novel development, the Deterministic Annealing Filter. We describe their b asic features and investigate their relationships with each other and with other popular estimators, in particular with the EM algorithm. The performa nce of the methods is optimized by means of several simulation experiments. We study the influence of the annealing schedule on the performance and sh ow that the application of advanced minimization methods is required in ord er to obtain reliable estimates of the track parameters. We compare the rel ative efficiencies and the computational costs of the four methods both und er ideal conditions and with noise. A final experiment under realistic cond itions focuses on the robustness of the proposed approach. (C) 1999 Elsevie r Science B.V. All rights reserved.