A new two-layer reinforcement learning approach the control of a 2DOF manipulator | IEEE Conference Publication | IEEE Xplore