Human-Inspired Decision Making for Swarm Robots
Abstract
Swarm robotics is the study on the development of interacting physical robots that are able todemonstrate intelligence as it is obs"erved in nature. Harnessing intelligence from nature,especially from humans, is important because it will enable not only unprecede""nted robotcapabilities but also more natural human-robot interaction, thus engendering more trust ofhumans on robots. Recent techn""ological advances provide a theoretical foundation for tacklingsome important issues in swarm robotics, such as efficiency and scal""ability. These advances,however, have only been based on utilizing information that is reliable for statistically optimaldecisions"" without addressing the unreliability of information and the associated decision risk,which are natural components of the human dec""ision making process; therefore, robots usingthese technologies are unreliable and not human-friendly. The goals of the project are" toovercome both theoretical and algorithmic challenges in developing risk-aware swarm robots withthe dynamic arrival of unreliabl"e information. Specifically, the research will focus on creating afoundational human-inspired decision making approach for swarm ro""bots to manage informationunreliability and decision risk, thus reducing the misuse of discrepant information that results insyste""m unreliability. In pursuit of this goal, the project will focus on three essential thrusts: (1)Reliability-based Information Manag""ement: this thrust will formally define, quantify, andimprove information reliability; (2) Risk-based Decision Making: this thrust"" will quantify,evaluate, and control decision risk subject to information unreliability; and, (3) ExperimentalValidation: this thr"ust will verify the proposed approach using actual robots conducting missionsthat are fundamental to the Department of Defense (DoD). The success of the project is expectedto fulfill DoD~s current needs to enable trustworthy robots for reduced human workload as well asDoD~s future needs to create next-generation swarm robots that can think and act like humans.The novelty of this interdisc"iplinary project is the synthesis of approaches from the social,management, data, and decision/control sciences to create a new hum""an-inspired decision makingapproach such that robots are human-friendly, reliable, and risk-controlled in the presence ofunreliabl"e information.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 09, 2017
- Source ID
- N000141712613
Entities
People
- Yongcan Cao
Organizations
- Office of Naval Research
- United States Navy
- University of Texas at San Antonio