TSAR: Trustworthy Search And Rescue uncrewed aerial vehicle

TSAR：值得信赖的搜救无人驾驶飞行器

• 批准号: EP/Z001102/1
• 负责人: Ardhendu Behera
• 金额: $ 26.26万
• 依托单位: Edge Hill University
• 依托单位国家: 英国
• 项目类别: Fellowship
• 财政年份: 2024
• 资助国家: 英国
• 起止时间: 2024 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FZ001102%2F1
• 关键词: TSAR; Trustworthy; Search; Rescue; uncrewed

Recent disasters around the globe, natural and manmade, claimed over a million lives and cost trillions of dollars in economic losses. Although disasters may not be prevented, rescue operations can considerably reduce the loss of life. Search And Rescue (SAR) is the foremost mission to be swiftly carried out in the aftermath of a disaster. First responders' search efforts in disaster situations, especially in contested areas (obstacles and foggy conditions) and war zones, are slow as they thoroughly examine the surroundings to avoid unnecessary debris shifts while protecting themselves. The existing state-of-the-art SAR Uncrewed Aerial Vehicles (UAVs) can only identify people in open areas using Artificial Intelligence (AI) algorithms. The proposed project, TSAR (Trustworthy SAR UAV), intends to develop an intelligent, trustworthy, reliable, and cost-effective SAR UAV that locates victims in contested areas and assesses their health. Such a UAV can expedite the search process and prioritize rescue missions for multiple casualties, enhancing successful rescue and saving lives. TSAR's research objectives: 1) Design a novel UAV for SAR operations considering vertical take-off and landing capability, high endurance and range, maneuverability, fault tolerance, and reliability factors; 2) Develop innovative computer vision algorithms to separate the contested areas from open areas, and adapt Voronoi tessellations to generate low-altitude search trajectories in the open areas to get a closer look at the contested areas; 3) Develop multimodal AI algorithms to fuse and process the data from the auditory and visual sensors to geo-locate victims; 4) Non-disruptive health check maneuvers providing maximum visual coverage of victims and their health status via a novel algorithm by analyzing heat signatures; and 5) TSAR's trustworthiness will be achieved by considering legal, ethical, system and human elements embedded through design, development and testing phases.

最近在地球仪发生的自然和人为灾害夺去了100多万人的生命，造成了数万亿美元的经济损失。虽然灾害可能无法预防，但救援行动可以大大减少生命损失。搜救（SAR）是灾难发生后迅速执行的首要使命。第一反应者在灾害情况下的搜索工作，特别是在有争议的地区（障碍物和雾的条件）和战区，是缓慢的，因为他们彻底检查周围环境，以避免不必要的碎片转移，同时保护自己。现有的最先进的SAR无人机（UAV）只能使用人工智能（AI）算法识别开放区域中的人。该项目名为TSAR（Trustworthy SAR UAV），旨在开发一种智能、可靠、可靠且具有成本效益的SAR UAV，用于在有争议的地区定位受害者并评估他们的健康状况。这样的无人机可以加快搜索过程，并优先考虑多名伤亡人员的救援任务，从而提高救援成功率并挽救生命。TSAR的研究目标：1）设计一种新型的无人机用于SAR作战，考虑垂直起降能力、高续航力和航程、机动性、容错性和可靠性等因素; 2）开发创新的计算机视觉算法，将有争议的区域与开放区域分开，并采用Voronoi细分法生成开放区域的低空搜索轨迹，以更近距离地观察有争议的区域; 3）开发多模式人工智能算法，融合和处理来自听觉和视觉传感器的数据，以地理定位受害者; 4）通过分析热特征，通过新算法提供最大限度的受害者及其健康状况的视觉覆盖，实现无中断的健康检查演习; 5）TSAR的可信度将通过考虑设计、开发和测试阶段嵌入的法律的、道德、系统和人的因素来实现。