Skill-Assist Safety and Intelligence Technology
Author(s) -
Suwoong Lee,
Yoji Yamada
Publication year - 2009
Publication title -
international journal of automation technology
Language(s) - English
Resource type - Journals
eISSN - 1883-8022
pISSN - 1881-7629
DOI - 10.20965/ijat.2009.p0643
Subject(s) - safety monitoring , computer science , controller (irrigation) , hidden markov model , focus (optics) , component (thermodynamics) , reliability engineering , real time computing , simulation , engineering , artificial intelligence , physics , microbiology and biotechnology , optics , agronomy , biology , thermodynamics
Skill-Assist for automobile manufacturing enables operators to move heavy component modules to target sites, playing a valuable role on production lines. Skill-Assist has high-powered actuators and operates in physical contact with users, so safety is a top priority. This paper describes safety technology developed for and implemented in Skill-Assist at the National Institute of Advanced Industrial Science and Technology, Japan (Skill-Assist@AIST). Risk was assessed for main causes of potentially hazardous events, which were projected to result from abnormal command signals generated by the controller, human error, and unauthorized access. In this paper, we focus on safety measures against abnormal command signals and human error, and introduce current safety technology for Skill-Assist@AIST. Highly reliable control includes a dual-channel controller and fail-safe fault-detection hardware (FSFDH) for ensuring functional safety through command signal monitoring. A reaching-gesture recognition (RGR) algorithm based on laser range sensor data and a hidden Markov model (HMM) predictively detect operator error that outlies predefined safe reaching.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom