Development of Metrics for the Fidelity Assessment of the Human-Machine Interface in Flight Simulation Applications

高保真建模和仿真是新飞机设计和开发过程中必不可少的工具。它还有助于开发可用于飞行员培训的环境，与实际操作相比降低成本，并提高安全性。飞机开发项目耗时长、费用高，而且大量使用建模和仿真技术。发展计划携带大量相关的风险直到飞行测试和随后的发展。如果设计问题在设计后期被发现，修复起来既耗时又昂贵：萨博鹰狮原型机在首次飞行着陆时遇到飞行控制问题，导致飞机坠毁。到1986年，F-16的研发成本上升了70亿美元，达到138亿美元，部分原因是飞行控制问题，导致几个月的延误，而最新的F35 A/B/C开发计划可能面临超过五年的延误。虽然在计算机处理能力（使更复杂的飞行模型能够实时运行）和改进的模拟器组件（更灵活的运动平台，更快的图形卡）方面继续取得进步，但如何利用这些工具来支持飞机开发计划仍然存在挑战。本研究的目的是减少模拟器组件对模拟器结果的（目前未知的）影响的风险和不确定性。当这些结果对飞机的“安全”至关重要，或者稍后需要付出高昂的代价（例如设计、飞行许可、首次飞行准备或训练质量）时，这一点至关重要，并将通过建立评估模拟器感知保真度的框架来实现，即模拟器组件对飞行员意见、评级和技能获取的影响。对于训练模拟器，国际标准如CS-FSTD(H)[1]规定了公差（预测保真度），该标准应用于生产具有足够“保真度”的飞行力学，以确保训练装置的实用性。CS-FSTD(H)还提供了模拟器其他单独组件的技术要求信息，例如视觉系统的视场，控制加载感觉。一旦组件安装到模拟器中，在其最终鉴定之前需要进行主观测试。标准中提供了一些关于在主观评价期间使用的方法的指导，但没有提供在这一过程中使用的客观度量或公差。标准中没有很好地捕获的是应该如何进行主观评估，应该使用什么度量来量化飞行和模拟之间的差异，因此可以接受的公差应该是什么，以便进行有效的训练。新的研究需要开发模拟器保真度评估的客观指标，以及模拟器主观评估的更稳健的方法。对于研究和工程模拟器，没有标准来定义模拟器的保真度水平，因此它们的效用；这将在拟议的活动中进行检查[拟议的问题将制定一个派生此类指标的框架]。在海军作战领域，飞机作战到舰船的安全范围是通过首级飞行试验（FOCFTs）确定的。目前，focft使用的是“活”资产，因此，focft昂贵、危险、耗时，而且其范围往往受到试验期间遇到的环境条件的限制。因此，有一种使用建模和模拟来提供证据的动力，以便为飞机清关方案制定安全案例；这意味着仿真环境的成熟度和有效性是至关重要的。BAE系统仿真部

High fidelity modelling and simulation is an essential tool in the design and development process for new aircraft. It also enables the development of an environment which can be used for pilot training, reducing costs compared with real-world operations and leads to improvements in safety. Aircraft development programmes are long and expensive and make extensive use of modelling and simulation. The development programmes carry a large amount of associated risk up to the point of flight testing and subsequent development. If problems with the design are discovered at late design stages they are time consuming and expensive to fix: The SAAB Gripen prototype experienced flight control problems during the landing of its first flight, leading to the loss of the aircraft. The F-16 research and development costs rose by $7 billion to $13.8 billion by 1986 in part due to flight control problems, leading to the months of delays, whilst the latest F35 A/B/C development programme is potentially facing more than five years of delays.Whilst there continues to be advances in computer processing power (enabling more complex flight models to run in real-time) and improved simulator components (more agile motion platforms, faster graphics cards) there are still challenges in how these tools can be utilised to support aircraft development programmes. The aim of the proposed study is to reduce the risk and uncertainty of the (presently unknown) influence of simulator components on the results derived from a simulator. This is paramount when those results are 'safety' critical for the aircraft or would be costly to put right later (e.g. design, clearance to fly, prep for first flight or training quality) and will be achieved by putting in place a framework for assessing simulator perceptual fidelity, i.e. the influence that the simulator components have on the pilot's opinion, ratings and skill acquisition.For training simulators, international standards such as CS-FSTD(H) [1] specify the tolerances (predictive fidelity) which should be used to produce a flight mechanics with sufficient "fidelity" to ensure the utility of the training device.CS-FSTD(H) also provides information on the technical requirements for other individual components of the simulator e.g. field of view for the visual system, control loading feel. Once the components have been installed into the simulator a subjective test is required prior to its final qualification. Some guidance on the methodology to be used during the subjective assessment is provided in the standards but no objective metrics or tolerances are provided for use in this process. What is not well captured in the standards is how the subjective assessment should be carried out, what metrics should be used to quantify to differences between flight and simulation and hence what the acceptable tolerances should be to enable effective training. New research is required to develop objective metrics for simulator fidelity assessment together with a more robust methodology for the subjective assessment of simulators. For research and engineering simulators, there are no standards to define the level of fidelity of the simulator, and hence their utility; this will be examined in the proposed activity [the proposed problem will develop a framework for deriving such metrics].In the domain of naval operations, the safety envelope for aircraft operations to ship is determined through First of Class Flight Trials (FOCFTs). FOCFTs are currently conducted using "live" assets and, as such, are expensive, hazardous, time consuming and their scope is often limited by the environmental conditions encountered during the trials. Hence, there is a drive to use modelling and simulation to provide the evidence in order to develop the safety case for aircraft clearance programmes; this means that the maturity and validation of the simulation environment is of paramount importance.BAE Systems Simulation Department