A keyframe based inertial-visual navigation system for a commercial UAS autopilot

用于商用无人机自动驾驶仪的基于关键帧的惯性视觉导航系统

• 批准号: 478199-2015
• 负责人: Liu, Hugh
• 金额: $ 1.82万
• 依托单位: University of Toronto
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=579363
• 关键词: keyframe; based; inertial; visual; navigation

Unmanned aerial vehicles (UAVs) or unmanned aerial systems (UAS) hold great promise for applications as diverse as natural resource monitoring, infrastructure inspection (pipeline, railways), agriculture, mineral exploration, journalism, and search & rescue operations. This is the most vibrant sector of the aerospace industry, growing more quickly than any other segment. In many potential applications, small scaled UAVs or UAS have the advantage of maneuverability, versatility and portability but face the challenge of limited power and payload. In addition, many systems rely on global navigation satellite systems (GNSS) for guidance, navigation and control in their autonomous operation, however it brings restrictions to regions where GNCC coverage is unreliable or even not available. On the other hand, compact and energy-efficient digital cameras are becoming widely used as the primary sensor as they are well suited to the stringent power and payload restrictions imposed by these small vehicles. It is natural to explore vision-based navigation as an alternative solution. In this proposed Engage project with Micropilot Inc., we aim at developing an inertial-visual assisted navigation system integrated with Micropilot's commercial autopilot. It implements a promising keyframe based simultaneous localization and mapping (SLAM) algorithm with a novel feature of gimbal controlled camera. In summary, the proposed image-based visual algorithm with gimbal controlled camera provides a novel and computationally efficient system that enables autonomous navigation in UAV flight missions. The technique will be integrated with the Micropilot autopilot and implemented to a Micropilot Helicopter UAV for technical demonstration.

无人机（UAV）或无人机系统（UAS）在自然资源监测、基础设施检查（管道、铁路）、农业、矿产勘探、新闻以及搜索和救援行动等各种应用中具有很大的前景。这是航空航天工业中最具活力的部门，增长速度超过任何其他部门。在许多潜在的应用中，小型无人机或UAS具有机动性、多功能性和便携性的优势，但面临着功率和有效载荷有限的挑战。此外，许多系统在自主运行中依赖全球导航卫星系统进行制导、导航和控制，但这给全球导航卫星系统覆盖不可靠甚至不可用的地区带来了限制。另一方面，紧凑和节能的数码相机 正被广泛用作主要传感器，因为它们非常适合于这些小型飞行器所施加的严格的功率和有效载荷限制。探索基于视觉的导航作为替代解决方案是很自然的。 在这个提议的与Micropilot公司合作的Engage项目中，我们的目标是开发一个惯性视觉辅助导航系统与Micropilot的商业自动驾驶仪集成。它实现了一个很有前途的关键帧的同时定位和地图（SLAM）算法的一个新的功能框架控制的相机。 总而言之，所提出的带有万向节控制相机的基于图像的视觉算法提供了一种新颖且计算效率高的系统，可以在无人机飞行任务中实现自主导航。该技术将与Micropilot自动驾驶仪集成，并应用于Micropilot直升机无人机进行技术演示。