Collision assessment for UAV

加拿大交通部、联邦航空局和国际民航组织认识到，无法看到(感觉)和避免(SAA)其他空中交通是阻碍无人驾驶飞行器(UAV)融入非隔离空域的主要障碍。最初的要求是无人机具有与人类飞行员类似的安全级别(ELO)，能够“看到”并避开交通。但Elos现在被一个更可量化的衡量标准所取代：目标安全水平(TLS)，即每飞行小时可接受的可能导致死亡的近距离碰撞的平均数量。我的研究重点是一种新的分析工具来确定TLS，该工具基于用于开发和验证TCAS的行业公认的统计(遭遇)模型。该方法是通过对大量可能的相遇进行快速模拟来统计确定TLS。一个新奇之处在于，该分析器是“技术不可知的”：它可以评估开发人员的早期SAA算法和监管机构的标准，而无需参考和在实施特定技术或采用特定标准之前进行评估。早期成果可能是对在人烟稀少地区执行无人机巡逻任务的风险进行科学量化，这可能会对加拿大政府关于北极主权的政策产生直接影响。从长远来看，通过与美国联邦航空局的承包商麻省理工学院林肯实验室合作，我们能够在起草美国联邦航空局SAA标准的早期阶段贡献“加拿大内容”，该标准很可能稍后被国际民航组织和加拿大交通部采用。在航空航天领域，SAA的发展被一种循环依赖所扼杀：监管机构正指望各行业提供投入，但许多行业在投资一项技术之前，正在等待一项批准的标准。这项研究将通过打破这一僵局来促进行业增长。它开创了一种方法，使行业能够在设备制造之前对技术进行评估，并使立法者能够在标准成为法律之前对其进行验证。由此产生的风险量化也将有助于保险业承保商用无人机服务：这是无人机行业增长的必然。

The inability to see (sense) and avoid (SAA) other air traffic has been recognized by Transport Canada, FAA and ICAO as a major impediment preventing the integration of unmanned aerial vehicles (UAVs) into non-segregated airspace. An initial requirement was for UAVs to have equivalent level of safety (ELOS) similar to a human pilot being able to "see" and avoid traffic. But ELOS is now replaced by a more quantifiable measure: targeted level of safety (TLS), the acceptable mean number of near collisions per flight hour which could result in fatalities. My research thrust is aimed at a novel analysis tool to determine TLS, based on an industry-accepted statistical (encounter) model used to develop and validate TCAS. The approach is to determine TLS statistically by fast simulation over a large number of probable encounters. A novelty is that the analyzer is "technology agnostic'': it can evaluate early-stage SAA algorithms from the developers and standards from the regulators, without reference to and prior to the implementation of a particular technology or adaptation of a certain standard. An early outcome could be a scientific quantification of risk (TLS) for UAV patrol duties over sparsely populated areas and this could have an immediate impact on the Government of Canada's policy on Arctic sovereignty. In longer term, by collaborating with MIT Lincoln Lab, a contractor to FAA, we are positioned to contribute "Canadian content", at an early stage, in the drafting of an FAA SAA standard, which will likely be later adopted by ICAO and Transport Canada. Within the aerospace sector, SAA development is stifled by a circular dependence: regulators are looking to the industries for input, but many industries are waiting for an approved standard before investing in a technology. This research will foster sector growth by breaking this deadlock. It pioneers a methodology which enables industry to assess a technology before equipment is manufactured, and legislators to validate a standard before it becomes law. The resulting quantification of risk will also help the insurance industry to underwrite commercial UAV services: a necessity for growth of the unmanned aircraft sector.