Comparative assessment of human machine interfaces for ROV guidance with different levels of secondary visual workload A. Vasilijevic, N. Miskovic, Z. Vukic Abstract— Majority of ROVs are underwater vehicles with relatively slow dynamics virtually providing a ROV pilot extra time to perform other tasks, such as inspections and arm operation. However, with many tasks performed simultaneously with flying, the relevant information is typically dispersed on a number of screens overloading the pilot’s visual channel. Surprisingly, there is very little research examining the unique human-factors problems associated with unmanned underwater vehicles. Use of audio display has been suggested as a means to reduce visual workload, to enhance situation awareness, and mitigate the visual and cognitive demands of contemporary ROV operations. Our research investigate the effects of sec- ondary visual tasks on operators workload and performance using standard visual navigation interface, augmented reality visual interface and audio interface. All experiments were performed on the state-of-the-art, real-time ROV simulator developed by Mobile & Marine Robotics Research Centre, University of Limerick and augmented reality system developed by Laboratory for Underwater Systems and Technologies, University of Zagreb. As expected, the results show that in no-load conditions visual guidance is better than the guidance- by-sound. By contrast, the effects of secondary visual load affect operators’ performance. The use of augmented reality paradigm and especially hearing, in the form of the auditory display, emerges as an important advantage. Improvement depends on a level of experience in using auditory guidance system. Practice has a major effect on performance, bringing us to the conclusion that there is a more room for improvement in using auditory interface. I. I NTRODUCTION The term remotely operated vehicle or ROV, in the mar- itime world, refers to a Unmanned Underwater Vehicle that is remotely operated by a human operator from the surface. The ROV is directly connected to the user interface through the umbilical cable (or tether) that transmits the power and communication. A contemporary ROV control room is stuffed with screens presenting everything from video streaming from multiple cameras to various data acquired from multiple ROV sub- systems. The information is exclusively presented visually. The pilot is often required to perform multiple tasks si- multaneously e.g. piloting, inspection, search and therefore enormous quantity of information [1], dispersed on different screens may easily overload the ROV pilots’ visual channel and prevent them from perceiving all important information related to the particular task. In ROV applications these issues, i.e. dispersion of relevant information, overloading of visual channel and operator multitasking, are recognized as a significant problem often resulting in failed missions or even mishaps. Surprisingly, there is very little research examining the unique human factors problems associated with unmanned underwater vehicles [2], [3]. Augmented Reality (AR) is a technology which, combin- ing a real-world scene with a virtual elements, change the way we receive information. All the relevant information that exists can suddenly become part of our decision making process. It embodies not only a usual (ordinary) visual blending of real and virtual worlds, but it can also merge capabilities of other human senses i.e. hearing or touch, unloading the operators’ visual channel and benefiting from additional advantages specific to that human sense. Humans use auditory modality for development and main- tenance of situation awareness in natural environments. We are able to determine the location of a sound source anywhere in the 360-degree space around us (even for those that are out of our field of view), monitor events at multiple locations simultaneously and to switch our focus of attention between sound sources at will [4]. Exploiting these human abilities it is reasonable to expect that operators situational awareness can be improved using spatial sound interface/display. To authors’ knowledge, there are not comprehensive stud- ies that compare performance of visual, AR or audio guid- ance methods. The aim of this paper is to compare, based on objective measures, ROV path following performance using different HMI’s: standard visual-navigation display, AR visual display and AR auditory display. We hypothesize that performance of visual navigation (standard or AR) for no or low workload scenarios will be superior to the audio navigation. The reasons for that are: spatial acuity of the visual channel is much better than that of the auditory channel [5], and humans use vision on a permanent basis for navigation, we are very well trained for visual navigation. We also hypothesize that applying additional visual and cognitive load the advantage of using AR paradigm and especially hearing, in a form of the auditory display shall come to the light. The objective of this paper is to present comparative test and analyze effectiveness of the different HMIs used. Hence, section 2 describes the methodology and a simulation platform used in the paper. Next, experimental results are provided in Section 3 and are discussed. Finally, a set of conclusions are provided. II. METHODOLOGY AND RESOURCES The nature of the experiments requires substantial number of experiments and test operators (subjects). The necessity of conducting the high cost in-field experiments was avoided by the use of the real-time ROV simulator expanded with visual and audio AR system. The ROV simulator was developed at the Mobile & Marine Robotics Research Centre (MMRRC),