首页 >

基于肌电和眼动信号的SVO-eVTOL操纵品质评估

李雨函1,张曙光2,吴义兵2   

  1. 1. 北京航空航天大学
    2. 北京航空航天大学交通科学与工程学院
  • 收稿日期:2024-09-30 修回日期:2024-11-12 出版日期:2024-11-14 发布日期:2024-11-14
  • 通讯作者: 张曙光
  • 基金资助:
    国家自然科学基金

Assessing Handling Qualities of SVO-eVTOL through EMG and Eye Data

  • Received:2024-09-30 Revised:2024-11-12 Online:2024-11-14 Published:2024-11-14

摘要: 随着城市空中交通(Urban Air Mobility, UAM)商业化运输的兴起,简化飞行操纵(Simplified Vehicle Operations, SVO)的概念被引入航空器设计,旨在简化操作流程以适应未来运输需求。然而,基于SVO设计的电动垂直起降飞行器(electric vertical take-off and landing, eVTOL)是否满足适航性条件,以及其操纵界面是否符合人机交互的友好性标准,目前尚存在不确定性。为了解决这一问题,对基于简化操纵设计的eVTOL进行了任务科目基元(Mission task elements, MTE)的实验,以评估其操纵品质。实验共招募了20名被试,通过库珀-哈珀量表收集了他们对eVTOL操纵品质的主观评分,并进行了半结构化访谈,同时记录了他们的眼动数据和肌电数据。本研究在结合主观评估和可解释的手工特征的基础上,提出了基于格拉米角场法和卷积神经网络的操纵品质评估模型。研究结果表明,不良的操纵界面设计显著影响了参与者的肌电信号和眼动信号。基于此结论设计的卷积神经网络模型能够以93.67%的准确率评估eVTOL的操纵品质。研究不仅为航空器操纵品质的客观评估提供了新的视角,而且为未来简化操纵eVTOL的设计提供了重要的改进依据。

关键词: 城市空中交通, 简化飞行操纵, 肌电, 眼动, 垂直起降飞行器, 操纵品质

Abstract: With the advent of commercial transportation in Urban Air Mobility (UAM), the concept of Simplified Vehicle Operations (SVO) has been integrated into aircraft design, aiming to streamline operational procedures to meet future transport de-mands. However, there is uncertainty regarding whether electric vertical take-off and landing (eVTOL) aircraft, which de-signed based on SVO, meet airworthiness criteria and whether their control interfaces adhere to ergonomic standards for user-friendliness. To address this issue, an experiment on Mission Task Elements (MTE) was conducted to assess the han-dling qualities of an SVO-based eVTOL. 20 participants were recruited for the experiment, during which their subjective ratings of handling qualities using Cooper-Harper Rating Scale, qualitative comments through semi-structured interviews, and electromyography (EMG) data and eye-tracking data were recorded. Additionally, a handling qualities assessment mod-el based on Gramian Angular Field (GAF) and 2D-convolutional neural networks (2D-CNN) was proposed. The results indi-cated that poor control interface design significantly affected the participants' EMG and eye-tracking signals. Benefiting from the spatio-temporal information provided by GAF images, the proposed 2D-CNN reached a classification accuracy of 93.67%. This study provides a new perspective for the objective assessment of eVTOL handling qualities and offers signif-icant guidance for the future design of SVO.

Key words: Urban Air Mobility, Simplified Vehicle Operation, electromyography, eye-tracking, electric vertical take-off and landing, handling qualities

中图分类号: