So-called chaotic storages get increasingly important in commercial use. Their high dynamics and resulting uncertainty about storage levels result in high requirements on logistic processes. The project presented in this paper combines methods to meet these requirements by the use of an autonomous stocktaking robot. It uses approaches from the field of cognitive inspired Artificial Intelligence, enabling the robot to act purposefully in an unknown environment. Even if the environment is constantly changing, the robot is able to acquire robust information about the current state of e.g., storage areas, their position and goods stored in them. The information gathered is of coarse granularity but is still be a valuable basis for the analysis and optimisation of intra-logistic pro-cesses.
