ABSTRACT
The DF plays an ever-increasing role in many biomedical, industrial automation, aerospace and environmental engineering processes and security systems in addition to many defence applications. As we have seen earlier the DF process gives (i) more spatial coverage of the object under observation/surveillance, (ii) provides redundancy of measurements – from similar and dissimilar sensors, (iii) enhances the robustness of the system’s performance – more types of measurements are available, and it is assured that at least one data set is always available, (iv) provides better prediction (with less uncertainty) of inferences – due to the use of enhanced information and (v) gives overall assured performance of the multi-sensor-integrated MSDF system. The complete process of DF involves several closely related disciplines [1,2]: (i) signal and/or image processing, (ii) computational/numerical techniques and algorithms, (iii) information-theoretic, statistical and probabilistic approaches/measures/metrics, (iv) sensors’ mathematical models, sensor control-management, and sensor configuration-optimisation, (v) soft computing techniques and (vi) system identification, state and parameter estimation. Several such techniques from various fields strengthen the analytical treatment, understanding and performance evaluation of the DF systems.
