Adaptive estimation scheme for radome error calibration

An adaptive Kalman filter bank with a built-in "learning" process for generating radome aberraton error slope estimates has gone through the preliminary design stages. This modern Kalman filter technique in the switching environment can recognize the rapid changing processes in the radome slope parameter which severely limits the conventional adaptive control and estimation approach. This unique adaptive feature involves calculating the critical a posteriori probabilities in a manner where the semi-Markov statistics of a random starting process are used to make the intermediate predictive step. The estimator results seem to follow the actual radome error changes very well with compensation for the seeker dynamic lag. Further improvement on the dynamic response time is needed. This scheme has been imbedded into the guidance filtering loop to evaluate the terminal guidance performance enhancements obtainable in the presence of a worst case radome error environment for the high-altitude, high-speed threat. Preliminary results indicate encouraging improvement in the miss-distance and acceleration command histories.