Algorithm for ranked assignments with applications to multiobject tracking

William L. Brogan

Journal of Guidance Control and Dynamics · 1989 · 27 citations · 6 references

Concepts

Abstract

A sequential algorithm is presented for obtaining a cost-ranked set of solutions to the assignment problem. Concurrently, a conservative bound can be calculated that indicates how many of the ranked set are better than other potential assignments that may have been missed. Applications to two important strategic defense initiative measurement-assignment problems, namely the cycle-to-cycle and the sensor-to-sensor problems, are demonstrated. The procedure is useful for initiating new object tracks as well as for assigning incoming measurements to established tracking filters. Knowledge of the ranked set of assignments, as opposed to a single optimum, is important because of measurement uncertainties. The prudent course may be to initiate several tentative tracks in certain close-call situations.

References

6