Fusion of heterogeneous sensors data

A configuration with heterogeneous sensors using different measurement approaches most likely overcome the problem of correlated measurement errors as they occur when employing a set of homogeneous sensors that suffer from the same problems. A heterogeneous approach requires sensor fusion algorithms that take the different uncertainties of the sensors into account. In this paper we elaborate two sensor fusion methods for this task. The first algorithm uses the estimated variance of each sensor measurement in order to find the optimal averaging weights. The second algorithm considers the covariances and thus provides a more sophisticated model at the cost of higher complexity in implementation and computation.