Reflective, 3-Dimensional Behavior Tree Approach to Vehicle Autonomy

Many of today's underwater vehicles have a limited set of pre-planned behaviors that are of varying utility. This is due, in part, to very low underwater communication rates and difficulties observing the vehicle's underwater behavior pattern. The video game industry is a multi-billion dollar enterprise constantly investing in high quality, high performance frameworks for controlling intelligent agents. One such framework is called Behavior Trees. This project proposes a novel autonomy framework enabling easily reconfigurable behaviors for both land based and underwater vehicles to discover and map acoustic nodes using a modular open systems approach based on behavior trees and action repositories.