Much of classical robotics is framed as a start-goal problem. These problems are aimed at guiding a robot in a partially or fully known environment to a pre-determined start and goal-ultimately becoming a graph-search problem However, these problems do not encompass all the scenarios encountered by robots. For instance, robots designed for search and rescue may need to plan in unknown environments where random variables dictate the reliability of robot sensors. Thus, arguably an equally important problem in robotics is exploration. In other words, given an unknown map and unreliable sensor readings, can a robot follow a somewhat efficient trajectory to build the correct map?
