 |
|
Multisensor data fusion is the process of integrating the output of disparate data sources into a single refined estimate. ISIS has produced world leading algorithms for modelling and tracking dynamic nonlinear unknown processes via a set of new algorithms, including modified ASMOD, local state linearisation, feedback linearisation, recurrent neural networks and hybrid mixture of export algorithms. These databased algorithms provide models from data of the sensors, formulate state space models, another instantiates Kalman-like filters to estimate states, which are then fused. In a multi-estimator algorithm to provide the overall unified state estimate. These algorithms have been benchmarked on some well known target tracking examples, as well as demonstrated on a helicopter flight guidance problem (with GKN, Westlands) and on a Marine Navigation System (with RACAL Research). |
