S. S. Virdi, Yong Thiang Ng, Yisi Liu, Kelvin Tan, Daniel Zhang
{"title":"利用眼动追踪数据评估海员的态势感知能力","authors":"S. S. Virdi, Yong Thiang Ng, Yisi Liu, Kelvin Tan, Daniel Zhang","doi":"10.1115/omae2022-80754","DOIUrl":null,"url":null,"abstract":"\n Situation Awareness (SA) is the perception of the current situation, comprehension of its meaning, and projection of what is going to happen in the near future. It is crucial for navigators to possess high SA in a navigational Bridge to mitigate the risk of human errors and to improve navigational safety. However, the current methodology to assess SA mainly rely on human experts, which might bring in potential problems such as bias, work overload, and it is also hard for the human experts to capture every fine detail of the behaviour of the seafarers being assessed. To overcome these, an objective and automated way to assess Situation Awareness is needed. In this work, eye-tracking data is used for the assessment of SA. With the eye tracking device, it is possible to localize where the navigator is looking at, and by applying computer vision with deep learning algorithm, the ongoing activity being executed by the navigator could be identified. In total 7 activities (using RADAR, ECDIS, checking of ship’s heading, and speed, checking data on Echo Sounder, and data related to ships maneuvering, and others) can be recognized which are used as indicators of SA. A set of training data was recorded using Tobii Pro Glasses 3 to train the deep learning algorithm and test the classification accuracy. To further verify the proposed eye-tracking based assessment, a preliminary experiment has been designed and carried out. Five subjects were recruited for data collection. A full-mission Advanced Navigation Research Simulator (ANRS) was used to provide scenarios for both training data collection and preliminary experiment. From the initial results, it shows that a recognition accuracy of > 99% can be achieved, which gives positive support to the eye-tracking based recognition. The analytics results using data from preliminary experiment also show great potential in using eye-tracking to assess SA of navigators. The proposed assessment could be used in both simulator and on-board and for multiple purposes such as performance evaluation, promotion to the next rank, and Continuing Professional Development.","PeriodicalId":23502,"journal":{"name":"Volume 1: Offshore Technology","volume":"202 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2022-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Assessment of Situation Awareness for Seafarers Using Eye-Tracking Data\",\"authors\":\"S. S. Virdi, Yong Thiang Ng, Yisi Liu, Kelvin Tan, Daniel Zhang\",\"doi\":\"10.1115/omae2022-80754\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\n Situation Awareness (SA) is the perception of the current situation, comprehension of its meaning, and projection of what is going to happen in the near future. It is crucial for navigators to possess high SA in a navigational Bridge to mitigate the risk of human errors and to improve navigational safety. However, the current methodology to assess SA mainly rely on human experts, which might bring in potential problems such as bias, work overload, and it is also hard for the human experts to capture every fine detail of the behaviour of the seafarers being assessed. To overcome these, an objective and automated way to assess Situation Awareness is needed. In this work, eye-tracking data is used for the assessment of SA. With the eye tracking device, it is possible to localize where the navigator is looking at, and by applying computer vision with deep learning algorithm, the ongoing activity being executed by the navigator could be identified. In total 7 activities (using RADAR, ECDIS, checking of ship’s heading, and speed, checking data on Echo Sounder, and data related to ships maneuvering, and others) can be recognized which are used as indicators of SA. A set of training data was recorded using Tobii Pro Glasses 3 to train the deep learning algorithm and test the classification accuracy. To further verify the proposed eye-tracking based assessment, a preliminary experiment has been designed and carried out. Five subjects were recruited for data collection. A full-mission Advanced Navigation Research Simulator (ANRS) was used to provide scenarios for both training data collection and preliminary experiment. From the initial results, it shows that a recognition accuracy of > 99% can be achieved, which gives positive support to the eye-tracking based recognition. The analytics results using data from preliminary experiment also show great potential in using eye-tracking to assess SA of navigators. The proposed assessment could be used in both simulator and on-board and for multiple purposes such as performance evaluation, promotion to the next rank, and Continuing Professional Development.\",\"PeriodicalId\":23502,\"journal\":{\"name\":\"Volume 1: Offshore Technology\",\"volume\":\"202 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-06-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Volume 1: Offshore Technology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1115/omae2022-80754\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Volume 1: Offshore Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1115/omae2022-80754","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
情境感知(SA)是对当前情境的感知,对其含义的理解,以及对近期将要发生的事情的预测。在航行桥梁中,具有较高的安全系数对于降低人为失误的风险,提高航行安全性至关重要。然而,目前评估SA的方法主要依赖于人类专家,这可能会带来偏见、工作过载等潜在问题,而且人类专家也很难捕捉到被评估海员行为的每一个细节。为了克服这些问题,需要一种客观和自动化的方法来评估态势感知。在这项工作中,眼动追踪数据被用于SA的评估。使用眼动追踪设备,可以定位导航员正在看的地方,并且通过应用具有深度学习算法的计算机视觉,可以识别导航员正在执行的活动。总共可以识别出7项活动(使用RADAR、ECDIS、检查船舶航向、航速、检查回声测深仪数据、船舶操纵相关数据等)作为SA的指标。使用Tobii Pro Glasses 3记录一组训练数据,训练深度学习算法并测试分类准确率。为了进一步验证提出的基于眼动追踪的评估方法,设计并进行了初步实验。招募5名受试者进行数据收集。利用全任务高级导航研究模拟器(ANRS)提供训练数据采集和初步实验场景。从初步结果来看,该方法的识别准确率可以达到> 99%,为基于眼动追踪的识别提供了积极的支持。使用初步实验数据的分析结果也显示了使用眼动追踪来评估导航员SA的巨大潜力。建议的评估可以在模拟器和在职中使用,并有多种目的,如绩效评估、晋升下一级和持续专业发展。
Assessment of Situation Awareness for Seafarers Using Eye-Tracking Data
Situation Awareness (SA) is the perception of the current situation, comprehension of its meaning, and projection of what is going to happen in the near future. It is crucial for navigators to possess high SA in a navigational Bridge to mitigate the risk of human errors and to improve navigational safety. However, the current methodology to assess SA mainly rely on human experts, which might bring in potential problems such as bias, work overload, and it is also hard for the human experts to capture every fine detail of the behaviour of the seafarers being assessed. To overcome these, an objective and automated way to assess Situation Awareness is needed. In this work, eye-tracking data is used for the assessment of SA. With the eye tracking device, it is possible to localize where the navigator is looking at, and by applying computer vision with deep learning algorithm, the ongoing activity being executed by the navigator could be identified. In total 7 activities (using RADAR, ECDIS, checking of ship’s heading, and speed, checking data on Echo Sounder, and data related to ships maneuvering, and others) can be recognized which are used as indicators of SA. A set of training data was recorded using Tobii Pro Glasses 3 to train the deep learning algorithm and test the classification accuracy. To further verify the proposed eye-tracking based assessment, a preliminary experiment has been designed and carried out. Five subjects were recruited for data collection. A full-mission Advanced Navigation Research Simulator (ANRS) was used to provide scenarios for both training data collection and preliminary experiment. From the initial results, it shows that a recognition accuracy of > 99% can be achieved, which gives positive support to the eye-tracking based recognition. The analytics results using data from preliminary experiment also show great potential in using eye-tracking to assess SA of navigators. The proposed assessment could be used in both simulator and on-board and for multiple purposes such as performance evaluation, promotion to the next rank, and Continuing Professional Development.