Awards issued prior to january 20, 2025, were funded under previous administrations and may not reflect the priorities and policies of the current administration.** if a tree falls in the forest,…

Awards issued prior to january 20, 2025, were funded under previous administrations and may not reflect the priorities and policies of the current administration.** if a tree falls in the forest, and there is no one to hear IT, does IT make a sound? While the question is decidedly philosophical, IT conveys our goal here perfectly. But we also want to know a few other things, e.g., how large the tree was, what kind IT was, and how much would IT have grown in the future. We want to be able to answer such questions for the entire forest. Doing so is the key to understanding the carbon ecosystem and its impact on climate change, to advance forest science and to provide decision making tools for the $325B global forestry industry. To do so, IT is necessary to (a) build detailed maps that can help understand a region as large as an entire forest and update these maps regularly to capture changes in the forest, (b) develop a fine-grained understanding of the shape and size of the vegetation in the forest, a catalogue of the various species of trees and shrubs in the forest, and (c) develop new methods to understand whether there is sufficient diversity in the forest, the health of the forest and its grown rate. We are a team of roboticists and machine learning researchers from the university of pennsylvania and forestry experts from the virginia polytechnic institute and state university. We envision a team of unmanned aerial vehicles (uavs/drones), each equipped with sensors like high-resolution cameras that can build a map a given region quickly as complex as a forest. This team of drones, aided by forest scientist can build maps of areas as large as 1000 acres by flying under the forest canopy. We envision that these drones will be able to produce fine-grained maps of small, user-chosen, areas of forest, e.g., catalogue the diversity of vegetation, identify if there are diseases or pest infectations. Such fine-grained mapping will be conducted using new methods that demonstrate a curious but busy BEE explorer, i.e., the drone will explore an unknown region and once IT obtains enough information from this region, automatically move on to a new region. This system will be demonstrated via experiments in the wharton state forests and virginia state forests for improved decision making in forestry, e.g., logging strategies in pine plantations, mapping individual tree locations, species and physiology.