The performance that can be achieved when tracking a missile that flies through a radar network is investigated using simulation models and filters. A network of radar stations that can measure monostatic and bistatic range and Doppler is used and the data from each station pair is fed into a Kalman filter. The performance of the tracking is evaluated and compared with the results using only momentaneous measurements. The comparison between momentaneous measurements and tracking shows that tracking significantly increases accuracy for position and velocity estimates. With low flying targets we find that a regular Kalman filter jumps between ambiguous solutions above and below ground, but with a few restrictions that problem is eliminated.