Multiple frequency line tracking using hidden Markov/A models

The application of hidden Markov models to the problem of tracking multiple frequency lines is considered. The tracker generates multiple discrete Viterbi tracks and automatically accounts for track initiation, termination, and ambiguous measurement. The proposed algorithm is different from the authors' earlier method in that all of the amplitude information contained in the envelopes of all the frequency cells over the gate of the tracker are employed. This means that the use of sensitive detection thresholds can be avoided, thus making the tracker more applicable in practice. Using the concept of a mixed track together with the transition probabilities of individual frequency lines and the difference between the amplitudes of signals, the difficult problem of data association is avoided.