A maneuvering target is difficult to follow because its turns are hard to i
dentify from noisy position measurements. It is shown here that, even at cl
ose range, an image-enhanced architecture is effective in a tracking applic
ation. The performance of a nonlinear, dual-sensor algorithm is contrasted
with sophisticated radar-only algorithms, and it is shown that sensor utili
zation can be expanded significantly with a dual-sensor estimation approach
.