Semester of Graduation

Fall 2024

Degree

Master of Science in Industrial Engineering (MSIE)

Department

Department of Mechanical & Industrial Engineering

Document Type

Thesis

Abstract

Automation is becoming increasingly common in manufacturing and assembly plants. The future lies in hybrid workspaces where the strengths of humans and robots complement each other, with robots excelling in precision, speed, and strength, and humans excelling in creativity, emotional intelligence, and complex decision-making. Collaborative robots can foster a more efficient and productive work environment by bridging the gap between human and machine capabilities. This study examines how semi-automated and automated modes impact human-robot collaboration, focusing on mental workload, trust, and task performance.

In this experiment, 58 participants performed a primary task alongside a collaborative robot assembling a miniature lamppost while taking turns in a shared workspace under different automation conditions. In the automated group, the robotic arm operated independently. In the semi-automated group, participants controlled the robot's movements with the robot requiring a signal from the human to initiate its task. The study included a secondary task where participants pressed a button in response to auditory stimuli to measure detection response time (DRT).

Physiological measures of Heart rate, Heart rate variability score, and LF/HF Ratio were measured, and self-reported measures of NASA TLX and Trust in Automation questionnaires were collected to measure Mental Workload and Trust ratings. Four hypotheses were tested, predicting that automated mode would result in higher workload, faster assembly completion, and slower response times, and lower HR and higher trust in semi-automated mode.

Results showed no significant differences in MWL or DRT between groups. A weak positive correlation between HR and DRT was observed in automated mode suggesting cognitive overload or disengagement. Temporal Demand subscale (TLX) showed a moderate negative correlation with response times in semi-automated mode suggesting faster reaction times under time pressure. Trust ratings were higher by 4.6% in semi-automated mode, reflecting participants' preference for user-paced control. Automated mode showed 22.3% faster assembly time, demonstrating the efficiency of pre-set automation pacing.

This study suggests that while automation can improve task efficiency, user control is crucial for maintaining cognitive engagement and trust. Balancing automation and human involvement is key to optimizing performance in human-robot collaboration, particularly in industrial settings where task complexity and time pressures can fluctuate.

Date

11-16-2024

Committee Chair

Ikuma, Laura

DOI

10.31390/gradschool_theses.6055

Download

Share

COinS