A Survey of Robotic Navigation and Manipulation with Physics Simulators in the Era of Embodied AI
A new survey paper explores the use of physics simulators in embodied AI research, focusing on their role in bridging the sim-to-real gap for robotic navigation and manipulation tasks. The paper analyzes simulator properties, features for these tasks, and hardware requirements. It also provides a resource guide for researchers, including benchmark datasets, metrics, simulation platforms, and methods to aid in tool selection while considering hardware constraints. AI