Learning Multisensor Confidence Using a Reward-and-Punishment Mechanism | IEEE Journals & Magazine | IEEE Xplore