Artificial neural network approach for reconstruction of event arrival-direction in wide-angle atmospheric cerenkov detector arrays: A feasibility study

A feasibility study has been carried out to assess the potential of an Arti ficial Neural Network (ANN) for determining the direction of incidence of a n Atmospheric Cerenkov Event (ACE) from the arrival-time information regist ered by a spaced-array of wide-angle Cerenkov detectors. The results obtain ed so far, using both, simulated and experimental data, indicate that a pro perly-trained net can yield a degree of accuracy which is comparable with w hat is achieved through the conventional chi (2)-minimization, wavefront-fi tting procedure.