Automated Emergency Landing System for Drones: SafeEYE Project
| Autor: | Jacob Naundrup Pedersen, Aitor Ramirez Gomez, O. M. Bektash, Anders la Cour-Harbo |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
021110 strategic
defence & security studies 020301 aerospace & aeronautics business.industry Process (engineering) Computer science 0211 other engineering and technologies Crash System safety 02 engineering and technology Service provider Track (rail transport) Drone 0203 mechanical engineering Aeronautics business Risk management Search and rescue |
| Zdroj: | Bektash, O M, Pedersen, J N, Gomez, A R & la Cour-Harbo, A 2020, Automated Emergency Landing System for Drones : SafeEYE Project . in 2020 International Conference on Unmanned Aircraft Systems (ICUAS) ., 9214073, IEEE, International Conference on Unmanned Aircraft Systems (ICUAS), pp. 1056-1064, 2020 International Conference on Unmanned Aircraft Systems (ICUAS), Athen, Greece, 01/09/2020 . https://doi.org/10.1109/ICUAS48674.2020.9214073 |
| DOI: | 10.1109/icuas48674.2020.9214073 |
| Popis: | Automated emergency response systems have been the focus for development of more reliable and robust safety systems, from simpler ones to the most complex. For drones, such systems can be designed to allow compliance standards, track safe places for landing and provide an easier development for operational process. Although many works have acknowledged a need for automated response to deal with the increasing drone safety concerns, the literature is still scarce on research incorporating drone operations. Given this outlook, this paper presents the SafeEYE project, which was initiated to develop and commercialise an automated emergency landing system for larger ($>$ 7 kg) drones. The system consists of a small embedded computer, mounted on a drone, that keeps track of safe places to land, or even crash, as well as the health state of the drone. When there is a failure condition, the device can monitor and detect issues using vibration data, select potential landing zones via a convolutional neural network framework and terminate the flight with the least probability of fatalities. This means a significantly reduced risk for automated, typically Beyond Visual Line of Sight, operations. Therefore, SafeEYE has the potential to become a safety enabler for many applications, including farming, inspection, transportation, search and rescue. With the risk mitigation ability, the project aims at achieving formal approval of the Danish authorities and abroad. SafeEYE is planned to be manufactured as a standalone unit, provided first through drone technology suppliers and later to service providers and manufacturers of autopilots. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|