A neural approach to a sensor fusion problem.
Valentina Colla, Mirko Sgarbi, Leonardo Maria Reyneri, Angelo Maria Sabatini · 1998
Our problem concerns the joint interpretation of UltraSonic and InfraRed measurements provided by a composite proximity sensor, in order to extract geometrical and morphological features of a at target. Neural networks are applied here both for modeling and for classification purposes. Physical knowledge of the sensing modalities helps in simplifying the network structure, in order to minimize its complexity, save computing time and make the system suitable for real-time applications.