Vision-aided Guidance and Navigation for Close Formation Flight

Abstract

Close formation flight can extend an unmanned aerial vehicle's (UAV) range and endurance by utilizing lift from a wingman's wake vortices and by autonomous midair refueling or recharging. The prohibitive challenge in each of these applications is the highly accurate and reliable relative positioning that is required to stationkeep in the wingman's wake and to dock, amid external disturbances. Global navigation satellite systems are well-suited to reliable absolute positioning, but they fall short for accurate relative positioning. This work proposes a relative positioning solution for UAV rendezvous and close formation flight that has been verified in multiple flight tests. A nonlinear estimation framework uses precise air-to-air measurements to correct onboard sensor measurements and produce an accurate relative state estimate that is resilient to intermittent relative measurement outages and degrades gracefully during extended outages. A guidance strategy compensates for wingman turn dynamics, acts explicitly on the estimated relative state, and is applicable to both rendezvous and formation flight. Ground testing showed a relative position estimate accuracy that is 2% of the separation distance, with successful detection and correspondence at up to 36 m. Autonomous close formation flight tests verified the relative positioning solution over extended periods, as close as two wingspans, in winds that were 30%-40% of the cruise airspeed, and at altitudes as low as 15 m. Root-mean-square relative position errors were 1.2 m horizontally and 0.44 m vertically during flights at the closest separation. C 2015 Wiley Periodicals, Inc.

VI-RPE: Visual-Inertial Relative Pose Estimation for Aerial Vehicles

“…MPE [10], tries all combinations for initialization and then runs a tracking algorithm to avoid the association step. As reported by Wilson et al [13], MPE tracking is not good enough to track one UAV from the other. In Section IV, we propose a novel multi-hypothesis tracker based on P3P and particle filtering that outperforms MPE and works well while tracking one UAV from another, widening the applicability of the method to more generic and outdoor scenarios, albeit posing greater robustness challenges to noise due to significantly more candidate detections per frame.…”

“…Autonomous close-formation flight: The recent work in [13] proposes a leader-follower control and pose estimation using two autonomous unmanned planes where the leader features infra-red markers and the follower has a camera embedded. They propose a simpler and faster marker association algorithm that takes in consideration an initial pose from the planes, showing large scale outdoor results.…”

Coordinated Path Following Control of Fixed-wing Unmanned Aerial Vehicles

“…We recommend the surveys in [1], [26], [27] to readers who are interested in the existing coordinated control results. Some methods have been validated by the formation flight of fixed-wing UAVs: in [28], a visionaided close formation flight of two UAVs was conducted based on leader-follower control; the work in [29] considered several formation design approaches, and performed field experiments of two UAVs with a combined controller; as a big advance in terms of the scale, autonomous flight of 50 fixed-wing platforms was presented in [30], but details of their control law were not provided.…”

Long-term real time object tracking based on multi-scale local correlation filtering and global re-detection

“…Visual object tracking, the problem of locating objects in a video sequence, is one of the central research topics in the field of computer vision [1] and Big Data Analytics (BDA). Recently more and more modern applications, such as precision guidance [2], intelligent surveillance [3], and auto-control system [4], etc., require object tracking result of high accuracy. According to the target changes in the process of tracking, object tracking can be divided into long-term tracking and short-term tracking.…”