Previous Abstract Return to Session D3 Next Abstract

Session D3: Aerial Vehicle Navigation

Multi-UAV Formation Geometries for Cooperative Navigation in GNSS-challenging Environments
Flavia Causa, Amedeo Rodi Vetrella, Giancarmine Fasano, Domenico Accardo, University of Naples Federico II, Italy
Location: Spyglass

Recently, an increasing effort has been concentrated on low altitude small Unmanned Aircraft Systems (UAS) applications such as small packet delivery, urban surveillance, infrastructure inspection, and three-dimensional mapping. In general, mission profiles require, or may require, the unmanned aircraft to fly in GNSS-challenging or denied environments such as natural or urban canyons, at least in a part of the mission. In these cases, autonomous flight is hindered by navigation issues, leading to significant difficulties in mission execution, or at least to the necessity of manual control in visual line of sight, which strongly limits UAS potential. Thus, a significant research effort is being carried out about safe autonomous navigation in challenging environments. The lack of reliable GNSS measurements is usually tacked by exploiting additional aiding information, e.g. vision aided navigation [1], positioning based on phone signals [2], and cooperative navigation [3].
Cooperative or networked navigation is a term used to describe an approach whereby a community of users exploits shared measurements and information exchange to the navigation advantage [4], [5]. Different strategies have been proposed, which are based on relative range/angles measurements [4], [6] or on the observations of common ground areas by onboard optical sensors [3], [7]. In this framework, the authors proposed in previous works [8], [9] an approach that exploits information broadcast and vision-based sensing and is based on the idea of carrying out complex missions by distributing functions among different flight platforms. In the considered scenarios, one or more vehicles flying in areas not susceptible to GNSS signal corruption, designed as “father”(s) UAV, are used to support autonomous navigation of a “son” UAV flying in challenging conditions, by broadcasting their positioning information and acting as features for vision-based tracking. Position, velocity, and attitude of the son UAV are estimated with an Extended Kalman Filter (EKF) that uses in the correction step the available (reliable) GNSS observables, estimates from magnetic and inertial sensors, eventual other information sources (e.g. vision-aided pose estimates), and cooperative measurements.
The proposed work follows this line of activity generalizing previous results, and has been developed within a research project aimed at a fully autonomous multi-UAV system, where autonomy refers to flight planning, navigation, and real time guidance.
In particular, the main contributions concern:
- an investigation of formation geometries able to adequately support cooperative navigation. This analysis is mainly focused on where to place the father UAV(s) in order to guarantee a given navigation performance for the son vehicle. This takes several aspects into account, such as: (coarse) three-dimensional structure of the environment, current GNSS satellites’ geometry, relative sensing measurements (range and/or angles), accuracy, and constraints (e.g., no line-of-sight obstructions, maximum inter-vehicle distances, illumination/background conditions for vision-based architectures), son navigation state and covariance (which depend on the embarked sensors). The outcomes of this analysis represent an important information to be used also for multi-UAV planning and guidance.
- a numerical performance assessment for different test cases, based on a purposely developed scalable simulation environment.
- an experimental analysis based on multi-UAV flight tests. In these tests, customized multi-rotors are flown in formation. They embark several sensors, ranging from optical cameras, to MEMS-based inertial units, to commercial GNSS receivers.
[1] G. Chowdhary, E. N. Johnson, D. Magree, A. Wu, and A. Shein, “GPS-denied Indoor and Outdoor Monocular Vision Aided Navigation and Control of Unmanned Aircraft,” J. F. Robot., vol. 30, no. 3, pp. 415–438, 2013.
[2] A. Lie, H. Mokhtarzadeh, P. Freeman, J. Larson, T. Layh, B. Hu, B. Taylor, D. Gerbe-Egziabher, P. Seiler, and G. Balas, “An Airborne Experimental Test Platform, from theory to flight,” Inside GNSS, pp. 40–47, 2014.
[3] Z. Zhu, S. Roumeliotis, J. Hesch, H. Park, and D. Venable, “Architecture for asymmetric collaborative navigation,” in Record - IEEE PLANS, Position Location and Navigation Symposium, 2012, pp. 777–782.
[4] H. Mokhtarzadeh and D. Gebre-Egziabher, “Performance of networked dead reckoning navigation system,” IEEE Trans. Aerosp. Electron. Syst., vol. 52, no. 5, pp. 2539–2553, 2016.
[5] A. J. Rutkowski, J. E. Barnes, and A. T. Smith, “Path planning for optimal cooperative navigation,” in Proceedings of the IEEE/ION Position, Location and Navigation Symposium, PLANS 2016, 2016, pp. 359–365.
[6] E. Cledat and D. A. Cucci, “Mapping GNSS restricted environment with a drone tandem and indirect position control,” in ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2017.
[7] V. Indelman, P. Gurfil, E. Rivlin, and H. Rotstein, “Real-Time Vision-Aided Localization and Navigation Based on Three-View Geometry,” IEEE Trans. Aerosp. Electron. Syst., vol. 48, no. 3, pp. 2239–2259, 2012.
[8] A. R. Vetrella, G. Fasano, and D. Accardo, “Cooperative navigation in GPS-challenging environments exploiting position broadcast and vision-based tracking,” in 2016 International Conference on Unmanned Aircraft Systems, ICUAS 2016, 2016, pp. 447–456.
[9] A. R. Vetrella, R. Opromolla, G. Fasano, D. Accardo, and M. Grassi, “Autonomous Flight in GPS-Challenging Environments Exploiting Multi-UAV Cooperation and Vision-aided Navigation,” in AIAA Information Systems-AIAA Infotech @ Aerospace, American Institute of Aeronautics and Astronautics, 2017.



Previous Abstract Return to Session D3 Next Abstract