Automatic Tracking Algorithm for Periodic Pulse Signal Based on Central Difference Kalman Filter

To address the problem that the traditional tracker was unable to maintain tracking when there were some missing pulses of periodic pulse signals, in this study, an automatic tracking algorithm for periodic pulse signals based on a central difference Kalman filter (CDKF) was proposed. Based on the actual situation in the application of sonar pulse reconnaissance, the target distance was relatively far with a limited bearing change rate, so using bearing and bearing change rate, two-dimensional linear motion and observation equation models were established. Then, an automatic start method of pulse tracking was presented. The method was based on pulse sameness discrimination. When the number of identical pulses exceeded the specified number, a tracker was started. Then, based on CDKF tracking, to predict and filter target bearing smoothly utilizing the established motion and observation equations, the automatic tracking procedure was presented. When there was a missing bearing observation in the input detection sequence, the predicted bearing value was used as a tracked one. When the signals could be detected again, the CDKF algorithm was operated using the relation between the predicted values and the current observations to maintain the tracking of the target. If the number of successive pulse deficiencies exceeded the specified, the tracking procedure was ended. Finally, simulation results of several situations with missing some stochastic pulses revealed that the effects of CDKF and KF had no significant difference on tracking even bearing change rate target. Moreover, although under linear motion and observation equation models, the effect of CDKF was better than KF on tracking dynamic bearing change rate targets. In the scenario of sonar pulse reconnaissance, adopting simple two-dimensional linear equations could ensure the effectiveness of tracking and avoid complicated calculations caused by nonlinear models.