Current Location: GT > COE > DCL Projects
Projects
Robotic sensor networks
Ayanna Howard
Sensor webs (or networks) have been shown to be a powerful tool for in-situ science applications ranging from earthquake forecasting to understanding climate change. These networks capitalize on their ability to deploy cheap nodes throughout a region of interest in order to gather information relevant for scientific analysis. Recently, there has been growing interest in mobile networks to deal with the limitations of static networks, including issues of network deployment, coverage, and fault tolerance. However, a number of issues still exist in deploying mobile sensor networks for science applications, including the effectiveness of adapting to the environment and to changing science requirements, balancing power usage, and selecting between communication and control strategies.
In this work, we employ a natural extension to the sensor web concept that enables controlled reconfiguration of sensor assets for fault-tolerant in-situ sampling. The main motivation behind our approach is to apply decentralized (i.e. local) control algorithms for network adaptation and deployment while establishing the global sensing capability required for science investigations. The integrated sensing platform combines hardware, in the form of communication/sensor devices, and simple mobility platforms for re-positioning sensor devices in response to changes in science demand, sensor failure, and/or communication dropout. This system of mobile sensors is conceptually described as a decentralized network of in-situ sensors to which scientists define objectives by identifying specific formations (or topologies) over specified regions that the sensor network should assume. To realize this science formation, software control is instituted for adaptive reconfiguration of the network that changes network topology, in effect establishing a self-adapting sensor network. This occurs in order to maintain the desired science-driven configuration in spite of changes in science demand, sensor failures or communication drop-outs. More...
UAV autopilots
Panagiotis Tsiotras
This project is an effort undertaken at the School of Aerospace Engineering at the Georgia Institute of Technology for the development of a low-cost Unmanned Aerial Vehicle (UAV) test-bed for educational purposes. The objective of this test-bed is to provide an avenue for the involvement of undergraduate students (primarily) and graduate students (secondarily) in UAV research. The complete design and development of all hardware interfaces of the UAV platform including the on-board autopilot. Based on flight test data a linear model has been developed for the lateral and longitudinal dynamics.
More...
Nanoparticle synthesis and process design
Martha Grover
Geodesic tractography segmentation
Allen Tannenbaum
The purpose of this work is to extract white matter tracts and volumetric fiber bundles from Diffusion-Weighted MRI (DW-MRI) in order to study clinically relevant outcomes related to these structures. The idea is to use directional information in an anisotropic energy functional based on the Finsler metric (remember that the Riemannian metric is one particular Finsler metric of interest) in order to extract the best geodesic (i.e. optimal) path, which we dub "the anchor tract", between two regions of interest. Subsequently, region-based active contours are used to segment the associated volumetric fiber bundle. Finally, statistics are computed along both the anchor tract and the volumetric fiber bundle which may be used to compare different clinical populations.
More...
Puppet choreography and automated marionettes
Magnus Egerstedt
Much like Hollywood and Broadway writers, puppeteers write scripts that decribe the elements that make up their performance. These elements include stage setups, characters, timing, and actions. Consequently, puppeteers use these detailed scripts to control the characters (puppets) in an expressive way while simultaneously adhereing to the structure and timing of the play.
Since the plays are described by discrete actions based on a clock mechanism, they may be modeled by a MDL, specifically the variation MDLp. Our research focuses on automating the process of translating puppet play specifications in MDLp to a sequence of valid puppet control modes. Not only are we interested in the specificaiton of these plays, we also wish to optimize the control sequence based on a model of a real puppet in order to create more realistic motion on the real platform.
Adaptive multiagent systems
Jeff Shamma
Multiagent systems involve a distributed collection of interacting decision making components, each
having access to local information and communication, each making local decisions, and each seeking to optimize local objectives
that may be in conflict with other components. An appeal of adaptive multiagent systems is the potential for dynamic reconfiguration in the face of an uncertain or evolving environment. An important distinguishing feature is that unlike traditional machine learning, each agent seeks to adapt in the presence of other adapting agents, and this feedback dependency significantly impacts achievable performance. This project takes a game theoretic approach to derive and analyze algorithms that enable adaptive multiagent systems to self configure into desirable operational configurations.