Evaluating Levels of Automation in Human–Robot Collaboration at Different Workload Levels

Dana Gutman,Yael Edan,Samuel Olatunji

doi:10.3390/app11167340

Abstract

This study explored how levels of automation (LOA) influence human robot collaboration when operating at different levels of workload. Two LOA modes were designed, implemented, and evaluated in an experimental collaborative assembly task setup for four levels of workload composed of a secondary task and task complexity. A user study conducted involving 80 participants was assessed through two constructs especially designed for the evaluation (quality of task execution and usability) and user preferences regarding the LOA modes. Results revealed that the quality of task execution and usability was better at high LOA for low workload. Most of participants also preferred high LOA when the workload increases. However, when complexity existed within the workload, most of the participants preferred the low LOA. The results reveal the benefits of high and low LOA in different workload situations. This study provides insights related to shared control designs and reveals the importance of considering different levels of workload as influenced by secondary tasks and task complexity when designing LOA in human–robot collaborations.

Highlights

Human–robot collaboration (HRC) involves one or more humans working with one or more robots to accomplish a certain task or a specific goal [1]
Workload can be influenced by task complexity as characterized in terms of the stimuli involved in the task for inputs, as well as the behavioral requirements the human operator should emit in order to achieve a specific level of performance [20]
The high levels of automation (LOA) produced better quality of task (QoT) execution compared to the low LOA

Summary

Introduction

Human–robot collaboration (HRC) involves one or more humans working with one or more robots to accomplish a certain task or a specific goal [1]. Workload addresses the actual and perceived amount of work that the human operator experiences as related to the effort invested in the task [14,15] It can be described in terms of the elements that constitute the cost of accomplishing the goal for the human operator in the HRC [16]. Workload can be influenced by task complexity as characterized in terms of the stimuli involved in the task for inputs, as well as the behavioral requirements the human operator should emit in order to achieve a specific level of performance [20]. The task properties include the component complexity—number of distinct actions that the human operator must execute or number of informational cues that should be processed (e.g., the number and type of subtasks to be managed, [22]); coordinative complexity—nature of relationships between task inputs and task products, the strength of these relationships as well as the sequencing of inputs (e.g., timing, frequency, intensity and location requirements [23]), and dynamic complexity—changes in the states of the environment which the human operator should adapt to [20,24]

Methods

Results

Conclusion