On-line learning optimal control using successive approximation techniques | IEEE Journals & Magazine | IEEE Xplore