The Kalman filter is globally accepted by estimation community and frequently applied in many real:time applications such as tracking, navigation, guidance, control process etc. The optimality of Kalman filter depends upon how accurate the mathematical models for actual dynamical system and measurement device are known. The performance of this filter also depends on tuning parameters such as process noise variance (0). and measurement noise variance (R). However, there can be13; situations when exact mathematical models may not be known or too difficult to model. In such cases, it is widely experienced that modelling errors are often compensated by overloading the tuning13; parameters of the filter by using trial and error method. But this becomes extra burden to filter13; designer and time-consuming task. To tackle such problems, Fuzzy logic can be one of the promising13; solutions that uses an intuitive experience based-approach for problems that are too difficult to model mathematically and for filters difficult to tune properly. This paper considers the combination of13; Fuzzy logic and Kalman filter that have traditionally been considered to be radically different. The13; former is considered heuristic and the latter statistical filtering. Two schemes such as Kalman filter13; and Fuzzy Kalman filter are applied for target tracking application and their performances evaluated13; using several numerical examples. The approach is relatively novel. Also comparison with one of the13; existing adaptive tuning algorithms is carried out. The performance is evaluated using certain error-13; based criteria.
展开▼
