Building brains for rooms: designing distributed software agents

Abstract

This paper argues that complex, embedded software agent systems are best constructed with parallel, layered architectures. These systems resemble Minskian Societies of Mind and Brooksian subsumption controllers for robots, and they demonstrate that complex behaviors can be had via the aggregates of relatively simple interacting agents. We illustrate this principle with a distributed software agent system that controls the behavior of our laboratory's Intelligent Room. Introduction This paper argues that software agent systems that interact with dynamic and complex worlds are best constructed with parallel, layered architectures. We draw on Brooks' subsumption architecture (Brooks, 1985) and Minsky's Society of Mind (Minsky, 1986) theory to dispel the notion that sophisticated and highly capable agent systems need elaborately complex and centralized control. Towards this end, we present an implemented system of software agents that forms the backbone of our laboratory's "Intelligent Ro...