Design of a robotic bio-sampler and localization improvement for underwater autonomous gliders.

摘要

This thesis comprises two parts, the first part presents the development of a robotic platform to function as a biological sampler compatible with the Slocum Underwater Autonomous Glider (UAG). The second part presents a localization algorithm to improve positioning estimation of the glider underwater.;The ocean is very critical to life on earth yet 95% of it still remains unexplored. Hence, scientists all over the world have been deeply interested in understanding all the features of the ocean. One such feature which still remains unclear is how a diverse bacterial community transitions between seasons in a coastal ecosystem and how this transition affects the global biogeochemical cycles. This is because of our inability to collect water sample at the right time and space in these ecosystems to resolve the processes influencing the microbiota. One of the reasons for this inability is the lack of a component capable of collecting and returning intact biomass to the laboratory for molecular ecology studies. To meet this requirement, the first part of this thesis aims at development of a robotic platform called the bio-sampler to address fundamental questions in marine ecology and to elucidate the mechanisms supporting the diversity of microorganisms in the ocean. Our aim is to have the bio-sampler installed in the science bay of the glider. Such a mobile platform is capable of in-situ sampling and preservation on a range of spatial scales. Using the bio-sampler we demonstrated autonomous filtration of samples and running our preservation process on them. We also conducted contamination and sample preservation tests to validate the functioning of this robotic platform. The results confirmed that the bio-sampler was able to perform sample preservation without carrying any water sample from previous sample to the next one. The results also confirmed that the bio-sampler did not cause any cross-contamination between samples.;In the second part of this thesis, we try to improve the localization of the Slocum glider. Underwater autonomous gliders such as the Slocum glider provide an effective platform for marine and coastal scientists for conducting exploration missions which may last several weeks or even months. However, localization of these gliders underwater is a challenging task based only on the sensors on-board these gliders. Also these gliders move slowly, with an average horizontal velocity of around 0.2 - 0.3 m/s and hence are vulnerable to ocean currents. When these gliders resurface, they receive GPS signals to identify their position. Since they mostly run underwater this makes it difficult to obtain accurate positioning of the glider. The new localization scheme is based upon the dynamic model of the glider fused with on-board sensor measurements like depths and yaw angles. The experimental results have shown that the new localization scheme improves the position estimation of the glider without using any new sensors apart from the ones which are already on the glider.