Adapting to the Dynamic Environment
Pervasive networks have to interact with and adapt to their environment, and have to cooperate in order to ful?ll some global task. The size of the system and its dynamics require solutions that are scalable and based on simple local rules, because there is no central authority available for control, adaptation and optimization. We believe that understanding the impact of environmental changes, mobile, heterogeneous nodes and local rules needs a well-founded mathematical basis, because experiments can only measure some of the many aspects of this combination of size and dynamics of the system and its environment. Based on this insight, we will address the challenges described above by working on the following tasks.
Here we deal with local distributed strategies to interact with the environment in order to execute given global tasks. Such tasks may be explorations resulting in a map of the environment, maintaining such maps under dynamically changing environments, assigning tasks to components of the system, etc. The challenges are that
- we are confronted with the overall restrictions of such systems, namely that their components are simple, heterogeneous, and have very limited knowledge about the overall system,
- the nodes have very limited and error prone means to communicate and to sense data from their environment, and
- they have to be able to distinguish “good” from “bad” nodes; e.g., we have to realize methods that allow nodes to decide whether they can trust each other. This becomes important as the networks of the sensors may change due to the environmental changes, and due to failures and movements of the sensors.