A Deep Reinforcement Learning Framework and Methodology for Reducing the Sim-to-Real Gap in ASV Navigation | IEEE Conference Publication | IEEE Xplore