Underwater Multi-Dimensional Path Planning for the Naval Postgraduate School Autonomous Underwater Vehicle II
Abstract
Traditionally path planning software has been developed in LISP or C. With the recent government mandate for the use of Ada, this thesis seeks to demonstrate the feasibility of using Ada for both path preplanning and real-time path replanning. Land vehicle path planning can be accomplished with two horizontal components. However, for autonomous underwater vehicles, the two horizontal components and a vertical component are required. Memory and computational speed restrictions dictate that special processing of the search space be conducted to optimize the time-space trade-off. In this research a four dimensional array of nodes (two horizontal components, one vertical component and one orientation component) is used to represent the search space. By use of an orientation component, the number of nodes that can be legally moved to is limited, in effect pruning the search space. Three search methods were investigated: the Tendril search, the Direction search and the Real-time A* search. The Tendril search is a wavefront, breadth-first search. The Direction search uses a vector field for path planning. The Real-time A* search uses the Tendril search to a specified search depth then applies a heuristic to determine the best path to expand upon.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1991
- Accession Number
- ADA246569
Entities
People
- Joseph Bonsignore Jr.
Organizations
- Naval Postgraduate School