A Reinfrocement Learning Approach to Online Learning in Control | IEEE Conference Publication | IEEE Xplore