ROV VR

Project Overview

Summary:

In collaboration with the Monterey Bay Aquarium Research Institute (MBARI) and the Olin School of Engineering, I evaluated the usability of a VR application for ROV operation, ROV-VR, in comparison to standard multimonitor desktop ROV controls.

Problem Statement:

Remotely Operated Vehicles (ROVs) are tethered, teleoperated underwater robots used for marine research in place of human divers in dangerous or difficult situations.

Operating ROVs can be challenging, as operators rely on data from cameras and sensors, where distortion effects and lack of depth cues can make performing subsea work difficult.

Team:

I worked with a team of 10 people, including developers, UX researchers, marine biologists and ROV operators.

Research Methods and Process

Role:

UX Researcher

Methods and Tools:

  • Mutlivariate within-subjects user study

  • Usability testing

  • Quantitative data analysis (SPSS)

Usability Testing

ROV-VR was evaluated on the HTC Vive Cosmos for both stereoscopic and monoscopic VR conditions, and a standard ROV operation room with a multimonitor desktop display was used for the non-VR condition.

After testing with each condition, the System Usability Scale (SUS) was used to evaluate the usability of the ROV controls using ROV-VR for stereo- and monoscopic conditions, as well as the multimonitor desktop condition.

User Study

A within-subjects user study was conducted to evaluate the efficacy of ROV-VR, in comparison to standard, multimonitor desktop ROV controls.

The following 3 conditions were evaluated:

  • Stereoscopic VR

  • Monoscopic VR

  • Standard multimonitor desktop

Efficacy was measured through evaluation of cognitive task load using the NASA TLX and task performance using timed midwater marine sample capture tasks for each of the conditions.

Additionally, presence was measured to evaluate the impact of engagement and flow during task performance, using the Slater-Usoh-Steed Presence Questionnaire (SUS-P).

Results

Quantitative Data

Of the 3 conditions, ROV operators performed tasks the fastest with the stereoscopic VR condition, reducing performance time by 47% in comparison to multimonitor desktop controls.

Cognitive task load, measured through several factors in the NASA TLX, found significant differences in perceived effort between the stereoscopic and desktop conditions, with stereoscopic VR controls resulting in lower perceived effort by participants.

Based on data from the SUS, the stereoscopic condition was also the most usable, followed by the monoscopic and the desktop display, respectively.

This effect was matched with respect to presense, with the SUS-P demonstrating greater presence in the stereoscopic condition, followed by monoscopic VR and then desktop display.

Insights and Takeaways

The stereoscopic VR condition improved task performance, reducing performance time by 47% in comparison to standard multimonitor desktop controls.

Greater usability and presence with the stereoscopic VR condition over the multimonitor desktop controls also suggests that the transition to stereoscopic VR controls may improve task performance, especially over longer periods of time.

This suggests that the implementation of stereoscopic VR controls for ROV operation could help reduce costs, where the average cost of a dive is approximately $3600/hr, and may help to improve conditions for ROV operators.


Findings from this study were presented and published to ACM VRST 2021, winning the "Best Paper" award