An Adaptive Planning Framework for Situation Assessment and Decision-Making on an Autonomous Ground Vehicle
Abstract
The primary contribution of this research is the design, implementation, and field testing of an Adaptive Planning Framework "APF" that can address the problem of autonomous operation in a complex, unstructured environment. It encapsulates a new and unique approach to dynamic situation assessment, behavior management, and decision-making. This research also included a literature review and development of a Reference Implementation. The thesis behind this research is that a well-organized, three-stage process of 1) understanding the current situation, 2) understanding the suitability and viability of the available behaviors in light of that situation, and 3) providing the capability to autonomously make and execute behavior-related decisions, all in real-time, provides new levels of intelligence to autonomous ground vehicles (AGV). This research was performed using the resources of the UF Center for Intelligent Machines and Robotics. This environment provided the ability to collaboratively explore engineering alternatives, create experimental software, and test it in a real-world setting, ultimately leading to the creation of the Reference Implementation. All this was aimed at validating the thesis of the research and producing a more robust APF, operationally proven in a representative physical environment. The Adaptive Planning Framework has been shown to be both a viable method for representing and managing complex, situation-dependent behavior on an AGV and a valuable contribution to researchers tasked with developing and fielding such a vehicle. The viability of the architecture and design was demonstrated by the development and testing of the Reference Implementation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2006
- Accession Number
- ADA474020
Entities
People
- Robert A. Touchton
Organizations
- University of Florida