Building blocks for adaptive modular sensing systems

Abstract

This paper describes the design and implementation of a set of building blocks for the construction of adaptive modular sensing systems. A hardware architecture is proposed which allows the development of a general set of modular components with embedded knowledge about their capabilities. These modular components are referred to as transducer interface modules (TIMs). A knowledge representation scheme enables individual TIMs to exchange information about their capabilities and form a collective identity. The overall system functionality is determined by the manner in which TIMs are connected. This functionality is supported by a distributed software architecture that allows configuration-specific algorithms to be automatically loaded into each module. These building blocks can be rapidly reconfigured to form modular systems that are expected to prove useful in many applications, including industrial control, inspection systems, mobile robotics, monitoring, and data acquisition.