Defining "Soldier Intent" in a Human–Robot Natural Language Interaction Context

Report No. ARL-TR-8195
Authors: Eric Holder
Date/Pages: October 2017; 62 pages
Abstract: The work described was conducted to support a US Army Research Laboratory Director's Strategic Initiative research program to examine intent as conveyed and understood between humans and machines and used to achieve cooperative goals (i.e., "teaming"). The goal of this report is to analyze the question of what intent means for this paradigm, considering the Army context(s) of intended use as well as users' expectations and a range of potential meanings and applications of intent-based behaviors. The results are a review of the relevant literature combined with input from the field and integrated into a consolidated framework. This framework identifies a range of types of intent behaviors scaled for complexity primarily on the need for increased interpretation and reasoning and the use of implicit/existing knowledge (not directly or immediately provided) to ensure understanding and compliance. This framework will provide a structure, a goal, and benchmarks for existing and future efforts.
Distribution: Approved for public release
  Download Report ( 0.791 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.
 

Last Update / Reviewed: October 1, 2017