An operator-centric mission planning environment to reduce mission complexity for heterogeneous unmanned systems

摘要

Over the past 40 years, the field of robotics has made a huge impact on the way the scientific community, industry, law enforcement, and the military conduct their operations, especially with the use of unmanned underwater vehicles (UUVs). UUVs are vehicles designed to execute specific tasks in an underwater environment, come in many shapes and sizes and are adaptable to many different applications. With the growing use of UUVs across the oil and gas industry, scientific ocean exploration, subsurface archeology, department of homeland security, and several branches of the military, planning a mission for a UUV, allocating tasks to individual vehicles, and re-planning certain objectives during a mission are all critical activities. Inefficiencies in any of these stages can hinder mission success. Even with current trends moving towards using multiple vehicles of different types (i.e., heterogeneous vehicles), problems encountered during mission planning and re-planning can be compounded. The research presented in this thesis is a step towards alleviating these problems by offering a set of tools to aid personnel during the mission planning and mission execution phases. These tools include a high-level UUV mission programming language (MPL), a supporting graphical interface, and a set of automated mission planning tools. The high-level MPL offers multi-vehicle support for groups of cooperating heterogeneous UUVs and the graphical interface allows personnel who are not familiar with the MPL to specify a mission without writing source code. The automated mission planning tools offer both single and multiple objective approaches for a group of heterogeneous UUVs utilizing heuristic, linear programming, and genetic based methods. Due to the harsh and unpredictable nature of UUV missions, there is also a dynamic mission re-planning tool that reassigns UUV tasks (either the tasks of a failed UUV or new tasks that arise) to other vehicles during mission execution. The end goal is to automate the process of both the mission planning phase and the mission re-planning phase. This will allow personnel to focus on the overall mission objective rather than worry about burdensome details such as assigning specific tasks to vehicles and the coordination of multiple vehicles, thus providing better utilization of vehicle resources and increase overall mission efficiency.