Joint Optimization of Detection and Tracking in Adaptive Radar Systems - Advances in Heuristic Signal Processing and Applications - page 134

Fig. 6.9 Algorithms on the performance plane: The steady-state RMS position error versus TLP.
Here, the prefixes “H”, “S”, and “D” correspond to “heuristic”, “static”, and “dynamic” ap-
proaches, respectively. Note that, as SNR increases, the performance of the algorithms gets closer
to the best achievable performance point, the Kalman filter with perfect data association

in the lowest SNR case (see Fig. 6.9 (a)), threshold optimization greatly improves the
performance.
If we zoom in this most critical lowest SNR case and consider other possi-
ble heuristic approaches whose desired false alarm probabilities are ranging from
P FA =

10 − 1 , we get the whole performance trajectory shown in
Fig. 6.10 . As shown, the DTOP schemes are the only algorithms whose perfor-
mances are located nearly at the lower left corner of the trade-off plane. Although
the static schemes have low steady-state RMS position error level, they may not
provide low TLP as shown in the figure. The dynamic optimization schemes have
better transient characteristic as compared to static ones. This is an important aspect
of DTOP schemes and leads to improved track loss performance . From the practical
point of view, it can be argued that having a lower low track loss percentage is more
critical than having a lower steady-state position error. So in that respect, DTOP
schemes seem to be more viable solutions in practice.

10 − 8 to P FA =