Collision analysis and safety evaluation using a collision model for the frontal robot-human impact

Jung Jun Park, Jae-Bok Song, Sami Haddadin

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

The safety analysis of human-robot collisions has recently drawn significant attention, as robots are increasingly used in human environments. In order to understand the potential injury a robot could cause in case of an impact, such incidents should be evaluated before designing a robot arm based on biomechanical safety criteria. In recent literature, such incidents have been investigated mostly by experimental crash-testing. However, experimental methods are expensive, and the design parameters of the robot arm are difficult to change instantly. In order to solve this issue, we propose a novel robot-human collision model consisting of a 6-degree-of-freedom mass-spring-damper system for impact analysis. Since the proposed robot-human consists of a head, neck, chest, and torso, the relative motion among these body parts can be analyzed. In this study, collision analysis of impacts to the head, neck, and chest at various collision speeds are conducted using the proposed collision model. Then, the degree of injury is estimated by using various biomechanical severity indices. The reliability of the proposed collision model is verified by comparing the obtained simulation results with experimental results from literature. Furthermore, the basic requirements for the design of safer robots are determined.

Original languageEnglish
Pages (from-to)1536-1550
Number of pages15
JournalRobotica
Volume33
Issue number7
DOIs
Publication statusPublished - 2015 Aug 29

Keywords

  • Collision analysis and model
  • Human injury analysis
  • Physical human-robot interaction
  • Safe robots

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Computer Science Applications
  • Software
  • Mathematics(all)

Fingerprint Dive into the research topics of 'Collision analysis and safety evaluation using a collision model for the frontal robot-human impact'. Together they form a unique fingerprint.

  • Cite this