12.3 Environment Uncertainty and Mapping

After reading Section 12.2, you may have already wondered what happens if the map is not given. This leads to a fascinating set of problems that are fundamental to robotics. If the state represents configuration, then the I-space allows tasks to be solved without knowing the exact configuration. If, however, the state also represents the environment, then the I-space allows tasks to be solved without even having a complete representation of the environment! This is obviously very powerful because building a representation of a robot's environment is very costly and subject to errors. Furthermore, it is likely to become quickly outdated.