Subventions et des contributions :

Subvention ou bourse octroyée s'appliquant à plus d'un exercice financier. (2017-2018 à 2022-2023)

This research develops a new multi model state estimation method for navigation of Micro Aerial Vehicles ( MAV s). MAVs are versatile flying platforms which are increasingly finding applications in industrial inspection, military surveillance and recreational domains. Their primary commercial applications are in the area of outdoor surveying and aerial videography. These applications almost exclusively employ GPS waypoint navigation technology and offline photogrammetry software, to autonomously generate accurate 3D maps and high resolution footage. GPS constrained inspection using MAVs is currently performed manually by radio control and often as single platform implementations. There is a growing need for developing systems and technology which can autonomously perform these routing inspection tasks while exploiting the operational efficiency of multiple MAV platforms. In this scenario an operator can specify an area of inspection of a structure, which is then surveyed autonomously by multiple MAV units while providing sensor telemetry feedback to the user. The capability relaxes the need of expert qualified pilots and waives the need of manual motion control of each platform.
Multi model estimation strategy is used in applications where multiple models are needed to fully predict the behavior of a system during its operating conditions. MAV systems operate in a range of scenarios with changing sensor availability and environmental disturbances. In order to achieve accurate state estimation during these changing conditions the proposed research will develop a novel multi model estimator design for MAV navigation. The estimator will be designed while focusing on two new sources of information for the navigation system. These include the dynamic model of the platform which models the aerodynamic interactions of the device, and the relative sensing capability of MAVs which exploits the availability of multiple vehicles in a given task. The study will examine new design strategies to achieve computational efficiency, robust performance, system stability, and relative sensing methods that achieve accurate coordination between platforms. The multi model navigation system has many GPS limited industrial applications where automated surveillance and mapping are required. These include ship structure inspection, oil platform underdeck and flare inspection, and disaster surveillance. Moreover, the expected theoretical developments of the proposed research would serve a broader scientific and industrial user base, which includes underwater navigation, intelligent transport systems and flight avionics. It is expected that this research will result in a number of publications with significant impact and training of 2 PhD and 3 Master’s students as part of the program.