Collaborative Exploration with Dynamically Configurable Sensing Agents

We present a novel approach that provides the necessary localization and mapping capabilities to allow a dynamic placement of independent sensing agents. The presented approach allows to build complex sensor configurations using simple agents that carry basic sensors, like e.g. low-resolution cameras, which can be combined to high precision and high accuracy systems through fusion of their information and through optimized sensor placement. We discuss the mathematical background and verify the resulting system on simulated and real world data. We compare our system to other similar systems in the field of collaborative localization and mapping.