ALBERT

Description: Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the Massachusetts Institute of Technology and the Woods Hole Oceanographic Institution September 2022.

Description: Automated information gathering allows exploration of environments where data is limited and gathering observations introduces risk, such as underwater and planetary exploration. Typically, exploration has been performed in service of a query, with a unique algorithm developed for each mission. Yet this approach does not allow scientists to respond to novel questions as they are raised. In this thesis, we develop a single approach for a broad range of adaptive sampling missions with risk and limited prior knowledge. To achieve this, we present contributions in planning adaptive missions in service of queries, and modeling multi-attribute environments. First, we define a query language suitable for specifying diverse goals in adaptive sampling. The language fully encompasses objectives from previous adaptive sampling approaches, and significantly extends the possible range of objectives. We prove that queries expressible in this language are not biased in a way that avoids information. We then describe a Monte Carlo tree search approach to plan for all queries in our language, using sample based objective estimators embedded within tree search. This approach outperforms methods that maximize information about all variables in hydrocarbon seep search and fire escape scenarios. Next, we show how to plan when the policy must bound risk as a function of reward. By solving approximating problems, we guarantee risk bounds on policies with large numbers of actions and continuous observations, ensuring that risks are only taken when justified by reward. Exploration is limited by the quality of the environment model, so we introduce Gaussian process models with directed acyclic structure to improve model accuracy under limited data. The addition of interpretable structure allows qualitative expert knowledge of the environment to be encoded through structure and parameter constraints. Since expert knowledge may be incomplete, we introduce efficient structure learning over structural models using A* search with bounding conflicts. By placing bounds on likelihood of substructures, we limit the number of structures that are trained, significantly accelerating search. Experiments modeling geographic data show that our model produces more accurate predictions than existing Gaussian process methods, and using bounds allows structure to be learned in 50% of the time.

Description: The work in this thesis was supported by the Exxon Mobil Corporation as part of the MIT Energy Initiative under the project ‘Autonomous System for Deep Sea Hydrocarbon Detection and Monitoring’, NASA’s PSTAR program under the project ‘Cooperative Exploration with Under-actuated Autonomous Vehicles in Hazardous Environments’, and the Vulcan Machine Learning Center for Impact under the project ‘Machine Learning Based Persistent Autonomous Underwater Scientific Studies’.

Description: Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Mechanical and Oceanographic Engineering at the Massachusetts Institute of Technology and the Woods Hole Oceanographic Institution February 2023.

Description: To predict and mitigate anthropogenic impacts on the ocean, we must understand the underlying systems that govern the ocean’s response to inputs (e.g. carbon dioxide, pollutants). Analytical models can be used to generate predictions and simulate intervention strategies, but they must be grounded with empirical observations. Unfortunately, there exists a technological gap: in situ instrumentation is often lacking or nonexistent for key parameters influenced by anthropogenic inputs. While discrete bottle samples can be collected and analyzed for these parameters, their limited spatiotemporal resolution constrains scientific inquiry. To help fill the technological gap, this dissertation presents the development of instrumentation for the ocean inorganic carbon system and microplastics. The first few chapters present the development process of CSPEC, a deep-sea laser spectrometer designed to measure the ocean carbon system through alternating measurements of the partial pressure of carbon dioxide (pCO2) and dissolved inorganic carbon (DIC). CSPEC uses tunable diode laser absorption spectroscopy (TDLAS) to measure the CO2 content of dissolved gas extracted via a membrane inlet. Chapter 2 derives membrane equilibration dynamics from first principles, thus enabling informed design decisions. The analytical results showed that cross-sensitivity to other dissolved gases can be introduced by the equilibration method, regardless of the specificity of the gas-side instrumentation. A new method, hybrid equilibration, leverages the membrane equilibration dynamics to improve time response without incurring cross-sensitivity. Chapter 3 presents POCO, a surface pCO2 instrument that employs TDLAS and a depth-compatible membrane inlet. Through laboratory and field-testing, POCO demonstrated that hybrid equilibration overcame the gas flux limitation of deep-sea membrane inlets. Chapter 4 presents CSPEC, which successfully mapped the carbon system near different hydrothermal features at 2000 m in Guaymas Basin, becoming one of the first DIC instruments field-tested at depth. Chapter 5 introduces impedance spectroscopy for quantifying microplastics directly in water. Microplastics were successfully counted, sized, and differentiated from biology in the laboratory: a step toward in situ quantification. The analytical tools and measurement systems presented in this dissertation represent a significant step towards increasing the spatiotemporal resolution of carbon system and microplastic measurements, thus enabling broader scientific inquiry in the future.

Description: This research was supported by the following funding sources: NSF Grant # OCE-1454067 NSF Grant # OCE-184-2053 Link Foundation Ocean Engineering and Instrumentation Ph.D. Fellowship MITMartin Family Society of Fellows for Sustainability Richard Saltonstall Charitable Foundation National Academies Keck Future Initiative (NAFKI DBS13)