The SAL interpreter for large-scale optimization in distributed control systems

Many sensor-rich control systems interact with spatially distributed physical environments. This paper describes the spatial aggregation language (SAL) interpreter, a programming environment to support data interpretation and control tasks for distributed physical systems. SAL provides a set of powerful, abstract components to represent and exploit spatial structures in distributed physical data at multiple levels of abstraction. The programming environment supports rapid prototyping of application programs and interactive manipulation of spatial structures. In comparison with existing tools, the SAL environment is centered on geometric and topological representations that encode and utilize domain-specific knowledge, such as metrics, adjacency relations, and equivalence predicates. We illustrate the use of SAL in decentralized control design for thermal regulation.