The optimum gate shape and threshold for target tracking

A validation gate for target tracking is discussed in this paper. We present the new gate size estimation method by maximizing the difference between the average of target detection within the gate and the average number of false measurements within the gate. We also prove that an ellipsoidal gate computed from Chi-square distribution tables is the optimal shape. In this new method, the gate volume is finite even if the probability of target detection is 1. These characteristics can be intuitively approved.