Evaluation of Human-Robot Collaboration Models for Fluent Operations in Industrial Tasks

TL;DR

This paper evaluates different human-robot collaboration models in an industrial assembly setting, demonstrating that an adaptive model significantly improves efficiency and reduces idle time compared to traditional timing and sensor-based models.

Contribution

It introduces an adaptive collaboration model and empirically compares its performance against traditional models in a real-time industrial task.

Findings

Adaptive model reduced total assembly time by 7%.

Adaptive model decreased total idle time by 60%.

Adaptive system outperformed timing and sensor-based models.

Abstract

In this study we evaluated human-robot collaboration models in an integrated human-robot operational system. An integrated work cell which includes a robotic arm working collaboratively with a human worker was specially designed for executing a real-time assembly task. Eighty industrial engineering students aged 22-27 participated in experiments in which timing and sensor based models were compared to an adaptive model developed within this framework. Performance measures included total assembly time and total idle time. The results showed conclusively that the adaptive system improved the examined parameters and provided an improvement of 7% in total assembly time and 60% in total idle time when compared to timing and sensory based models.