Neural-network-based reinforcement learning controller for nonlinear systems with non-symmetric dead-zone inputs | IEEE Conference Publication | IEEE Xplore