An efficient strategy for data exchange in multi-robot mapping under underwater communication constraints

The online generation of underwater image maps or mosaicking is of high interest for underwater robots, e.g., for autonomous navigation, exploration, or object detection. Here, a cooperative approach is presented that addresses the particular challenges of the severe constraints on communication bandwidth in the underwater domain. Concretely, a special update strategy for a cooperatively maintained pose graph as basis for Simultaneous Localization and Mapping (SLAM) is introduced. The strategy tries to transmit the most relevant information within the limits of the communication bandwidth to maximize the quality of the cooperative map. It is shown in experiments with simulations based on real world data that the strategy leads to near optimal results while obeying the severe bandwidth constraints of realistic underwater communication.