A comparison of aircraft trajectory predictors for target tracking using simulated data

Three types of trajectory predictors used for target tracking are discussed and numerically compared in this paper. Included are predictors that model future acceleration as zero, constant, or time varying. In the class of time varying acceleration models, three variants are discussed. The first, the coordinated turn predictor, constrains future motion to one time-invariant maneuver plane. The motion within this maneuver plane is governed by two nonlinear differential equations. The second, the circular arc predictor, describes future motion as being circular in a time-invariant maneuver plane. And the third, the six degree-of-freedom (6-DOF) predictor, allows for future motion to result from the solution to differential equations for both translational and rotational degrees of freedom. The performances of three predictors are compared numerically on five different fixed-wing aircraft maneuvers. The maneuvers are typical of fighter aircraft in low altitude tactical situations and include a 5g pull-up, 5g right turns, and a more complicated S-turn maneuver. The results suggest that the coordinated turn predictor offers generally superior performance. The circular arc predictor, while not better than the coordinated turn predictor, outperforms the 6-DOF predictor.