NRI: An Ecologically Curious Robot for Monitoring Coral Reef Biodiversity

This grant will fund research that enables monitoring and management of coral reefs, which are critical biodiversity hotspots that provide direct economic benefits for subsistence and commercial fishing, tourism, and through coastal protection, thereby promoting the progress of science and advancing the national prosperity and welfare. Coral reefs worldwide are under threat from anthropogenic disturbances including climate change and associated ocean acidification.

New tools are needed to scale up monitoring of coral reefs to understand reef ecosystems, rapidly assess biodiversity, and measure the efficacy of interventions. This interdisciplinary project will address this need by creating an autonomous robotic system that can navigate a complex ecosystem and intelligently sample its environment to estimate local biodiversity and ecosystem health.

While the project will focus on an underwater system for monitoring coral reefs, the approach for estimating biodiversity is directly applicable to monitoring of other complex and threatened ecosystems such as rainforests, or for seeking out and mapping biodiversity in other remote habitats. A partnership with a minority-serving institution in the US Virgin Islands, the site for planned field expeditions, as well as participation in the Woods Hole Institute Summer Student and Minority Fellowship program, will help broaden participation in STEM of students from currently underrepresented groups and the institutions that serve them.

This research aims to make fundamental contributions to the integration of robotic technologies in an underwater autonomous vehicle that is capable of intelligent path planning, decision-making, and locomotion in a complex environment using high-dimensional information gained from acoustic and visual measurements. It will achieve this outcome by developing a novel seafloor hopping robot that is able to alternate between long stationary periods of observing its environment, while minimizing disturbances to animal life and conserving its energy budget, with short bursts of swimming activity from one landing spot to the next.

The project will develop a novel solution to the informative path planning problem particular to categorical data associated with observations of different types of habitat, plants, and animals. It will model this problem as a partially observable Markov decision process and will explore new approaches for computing expected reward rollouts in terms of the error between the estimated and true biodiversity of the environment, enabled by deep learning techniques and Bayesian nonparametrics.

Field expeditions will include validation of visual reef survey algorithms for habitat classification, collection of training data for machine learning algorithms, and tests of the path planning algorithm for locating biodiversity hotspots and long mission durations.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.