Middleware Navigation

I develop middleware that that allows information to flow between sampling platforms deployed in the water to flow into simulations of the water's dynamics, and to GCCS, the Glider Coordinated Control System, to provide navigation support for underwater vehicles. I am adapting GCCS to provide navigation support for the EcoMapper AUV. I'm very interrested in improving AUV guidance using the ROMS Ocean Model. My introduction to robotics consisted of working with Slocum Underwater gliders at the Rutgers University Coastal Ocean Observation Laboratory . This experience kindled my interest sampling the ocean, underactuated control in non-linear flows with large environmental uncertainty, building underwater robots and teaching. During my first year in Savannah, I lead GTSR (then GTS-AUV) to complete the 'Alpha' ROV, taught high school students participating in the Explorer post problem and served as head judge in the regional Lego robotics tournament. I am primarily focused on simultaneous localization and mapping of static and dynamic geophysical fields by submerged AUVs and AUV networks. At present, I am developing autonomous navigation algorithms for YSI's EcoMapper vehicle. To enable the development and testing of autonomous navigation algorithms for the EcoMapper, and AUVs, ROVS and ASVs developed on campus, I am designing and deploying a facility for conducting aquatic robotics experiments on campus. To support the research group's predictive sampling initiatives, I am developing simple in house implementations of the ROMS Ocean Model, which we will allow us to predict flow and assess environmental uncertainty. Access to water navigable by aquatic robots often limits the researchers' opportunity to develop and validate underwater localization and mapping methods. Due to Georgia Tech. Savannah's geographical location, our campus features on-site wetland area that we are developing for use by aquatic robots and networks. This testing area will consist of a dock, experimental water volume (up to 6m depth), a WI-FI network for surface communications and an acoustic global localization system. This resource will provide a safe, secluded environment that is under only natural forcing. During July 2010, a bathymetric survey of the lake was conducted, which covered most of our wetland. The resulting dataset was used to create a high resolution grid of the experimental area. We will, over time, complete and improve the accuracy of this map. We hope to expand our data coverage to include the missing region by exploring it with an Autonomous Vehicle. Last year, the LAMON research group obtained an EcoMapper AUV. We are currently developing methods for adjusting the vehicle's underwater position estimate based on sonar measurements, a stochastic model of the measurement process and a prior knowledge consisting of bathymetric survey data. To support predictive sampling experiments both on site and in the open ocean, I am developing implantations the ROMS Ocean Model. The Regional Ocean Modeling System will allow us to predict the environmental flow based on satellite altimetry and in-situ observations. Flow estimates may be used for generating control input for autonomous vehicles such as predicting the source of a chemical contaminant given in situ measurements of the contaminant. I also support the LAMON by administering the LAMON server, website, blog, and provide webhosting support for GTSR and by helping to 'make anything to work'. (see summary)


Figure 1. Justin Shapiro


Figure 2.