COMSTAR
Funding: DoD-Navair STTR grant via 21-st Century Systems Inc. (2006-2008)
In this project. we are building techniques that enable teams of autonomous robots to collaboratively perform complex tasks in an unknown environment. We use mini-robots like the e-puck that are not very expensive to build. However, these robots are constrained in the amount of on-board memory and the accuracy and sophistication of their sensors (e.g., camera, range-finding sensors). The limited capabilities on a robot makes it difficult for it to complete a 'complex' task on its own using only its on-board resources. Therefore, to complete a task, multiple robots must coordinate their actions with each other, so that they can perform the task collaboratively. Tasks can appear dynamically and must be discovered by the robots by exploring the environment. Finally, the coordination between the robots must be done in a completely distributed manner, with little or no interaction with a centralized base station. We have developed novel algorithms for distributed multi-robot task allocation, dynamic path planning and distributed exploration/coverage by the robot teams. We have tested these algorithms both within a robot simulator called Webots. and, currently testing them on physical e-puck robots. The details of our techniques can be found in our papers on the publications page.
Students
Current: Ke Cheng (Ph.D), Janyl Jumadinova (MS), John Petersen (MS)
Past: Matthew Hoeing (MS), Stephen O'Hara (MS)
MASSES (Multi-Agent Swarmed Space Exploration System)
Funding: NASA Nebraska EPSCoR (2007-2008)
The main focus of this project is to build a hierarchically controlled swarm of mobile robots capable of exploring extra-terrestrial bodies, using the Webots simulation environment. We have developed techniques for task discovery and allocation between the swarm units. Currently, we are investigating techniques for constellation formation and swarm taxis the swarm units to enable the movement of the swarm without disconnecting or disintegrating itself.
Watch a demo video from this project here.
Students
Past: Ross Bell (BS)DynaPAC (DYNAamic Pricing algorithms for Agent Coordination)
In this project, we are developing novel techniques for multi-attribute dynamic pricing by sellers in a market economy. Previous work in this area restricts itself to pricing good in a market economy, only along a single attribute, and, assumes that sellers know a priori different buyer parameters, such as buyer demand, number of buyers, buyer strategies, etc. We are addressing the harder problem of developing appropriate pricing strategies for sellers along multiple attributes of an item, without providing any a priori information about buyers to sellers. Our techniques have resulted in improved/accurate pricing resulting in higher profits (utilities) to sellers. We are also applying our techniques to multi-robot coordination.
Students
Current: Janyl Jumadinova (MS)
Past: Matthew Hoeing (MS), Yoshitsugu Hashomoto (MS)
PeerNets
In this project, we are developing techniques to enable rapid search in unstructured peer-to-peer networks. We have mainly developed algorithms based on emergent properties of artificial insect societies to enable rapid and dynamic network (re)-construction and fast searching.Students
Past: Erik Antonson (MS), Boonyot P. (MS)