AN ADAPTIVE EXTENDED KALMAN FILTER FOR ANGLES-ONLY TRACKING/INTERCEPT PROBLEMS

When angles-only measurements are used to estimate a target's ''state'' in tracking/intercept problems, closing geometries frequently encountered exhibit poor observability of some parameters which can result in divergence of the filter. In order to help mitigate the effects of poor observability, an adaptive extended Kalman filter has been developed to improve the accuracy of the estimates of relative position and velocity and the acceleration of a target using the angles-only information. This adaptive filter is based upon ridge-type estimation techniques and includes the computation of a biasing parameter, which provides a measure of the overall error in the solution and is used as input into the process noise matrix required in tuning the extended Kalman filter. The use of this adaptive filter will allow the tuning parameters to be adjusted each time an observation is processed. This adaptive extended Kalman filter is applied to a six degree-of-freedom homing intercept problem where angular measurements of the target's azimuth and elevation, relative to the interceptor, are the only measurements provided. Results are obtained for a number of different engagement geometries and are compared to the results obtained using both an untuned extended Kalman filter and a manually-tuned extended Kalman filter. Comparisons are made based upon the miss distance of the engagement as well as overall filter performance. Results show that the adaptive filter is similar in accuracy to that of the manually-tuned filter but is more responsive to errors in the estimates.