Learn More
We introduce a framework, called " physicomimetics, " that provides distributed control of large collections of mobile physical agents in sensor networks. The agents sense and react to virtual forces, which are motivated by natural physics laws. Thus, physicomimetics is founded upon solid scientific principles. Furthermore, this framework provides an(More)
In this paper, we explore the use of genetic algorithms (GAs) as a key element in the design and implementation of robust concept learning systems. We describe and evaluate a GA-based system called GABIL that continually learns and refines concept classification rules from its interaction with the environment. The use of GAs is motivated by recent studies(More)
— This paper presents an application of a physics-based framework for distributed control of autonomous vehicles. The autonomous swarm uses local information to self-organize into dynamic sensing and computation grids during localization of the source of a toxic plume. Using physics of fluid flow we develop a new plume-tracing algorithm, and then use(More)
The problem of designing and rening task-level strategies in an embedded multiagent setting is an important unsolved question. To address this problem, we have developed a multistrategy system that combines two learning methods: operationalization of high-level advice provided by a human and incremental renement by a genetic algorithm. The rst method(More)
This paper presents a rigorous evaluation of a novel, distributed chemical plume tracing algorithm. The algorithm is a combination of the best aspects of the two most popular predecessors for this task. Furthermore, it is based on solid, formal principles from the field of fluid mechanics. The algorithm is applied by a network of mobile sensing agents(More)