DISTRIBUTED UAV-SWARM-BASED REAL-TIME GEOMATIC DATA COLLECTION UNDER DYNAMICALLY CHANGING RESOLUTION REQUIREMENTS

The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2/W6, 2017 International Conference on Unmanned Aerial Vehicles in Geomatics, 4–7 September 2017, Bonn, Germany DISTRIBUTED UAV-SWARM-BASED REAL-TIME GEOMATIC DATA COLLECTION UNDER DYNAMICALLY CHANGING RESOLUTION REQUIREMENTS Miguel Almeidaa , Hanno Hildmannb and Gürkan Solmazc a b NEC Laboratories Europe, Kurfürsten-Anlage 36, D-69115 Heidelberg, Germany, Universidad Carlos III de Madrid (UC3M), Av. Universidad, 30 - 28911 Leganés - Spain, c NEC Laboratories Europe, Kurfürsten-Anlage 36, D-69115 Heidelberg, Germany, KEY WORDS: Self-organization, adaptive behaviour, swarm algorithms, distributed sensing, autonomous decision making ABSTRACT: Unmanned Aerial Vehicles (UAVs) have been used for reconnaissance and surveillance missions as far back as the Vietnam War, but with the recent rapid increase in autonomy, precision and performance capabilities - and due to the massive reduction in cost and size - UAVs have become pervasive products, available and affordable for the general public. The use cases for UAVs are in the areas of disaster recovery, environmental mapping & protection and increasingly also as extended eyes and ears of civil security forces such as fire-fighters and emergency response units. In this paper we present a swarm algorithm that enables a fleet of autonomous UAVs to collectively perform sensing tasks related to environmental and rescue operations and to dynamically adapt to e.g. changing resolution requirements. We discuss the hardware used to build our own drones and the settings under which we validate the proposed approach. 1. INTRODUCTION Autonomously operating Unmanned Aerial Vehicles (UAVs) have become a major technology in the past decade (though the U.S. military has been using UAVs for operations as far back as the Vietnam War (Broad, 1981)). Due to their low cost and high availability, airborne devices have received much interest from the private consumer, the research community, the industry and the military alike. So-called drones are referred to in the literature as UAVs (Schneider, 2014), UASs (Unmanned Aerial Systems) (Coopmans, 2014), RPAs (Remotely Piloted Aircrafts) (Marcus, 2014) or ROAs (Remotely Operated Aircrafts) (Ogan, 2014). There are many ways to classify UAVs (Malone et al., 2013). Drone types and capabilities are probably as numerous as the variations on the missions where drones are involved. Some of the larger (fixed wing) UAVs can operate for hours circling over certain areas, e.g., areas contaminated by highly hazardous materials (areas where a manned mission is too dangerous to human life or where providing adequate security for the human operator is simply too costly) (Malone et al., 2013), while the typical flight time of a commercial and publicly available (and affordable) offthe-shelf quadrotor is about 15 to 20 minutes (Erdelj et al., 2017). State of the art UAVs can remain airborne for prolonged periods of time (two weeks (Garber, 2014) or longer) and performance values improve significantly every year (Pauner et al., 2015). While a lot of research was undertaken regarding the autonomous landing of UAVs in general, few studies have been directed at marine environments where the challenges are more complex or circumstances are more dramatic compared to land or indoor environments. Recently, quadrotor UAVs have been landed with reasonable accuracy on swimming objects and under outdoor conditions (Mendona et al., 2016). The abilities to perform complex maneuvers and to operate as large collectives of units, especially for efficient situational awareness (Erdelj et al., 2017), are going to continue to increase rapidly over the next years. 1.1 Relevant Application Areas The United States Office of the Secretary of Defense (OSD) identifies over twenty UAV mission types, ranging from intelligence, surveillance (Giyenko and Cho, 2016a), reconnaissance missions (Broad, 1981), force protection, firefighting, electronic warfare to communication nodes and others (Malone et al., 2013). 1.1.1 Civil use mission UAVs have recently been used for civil supply, inspection and various search and rescue operations (Cummings et al., 2014). For example, UAVs are used for water management and biofuel production (Coopmans, 2014) or to monitor areas for wild-life protection (Schneider, 2014). For example, the World Wildlife Fund (WWF) controls poaching and illegal wildlife trade (Goodyer, 2013) and in some of Africa’s national parks engages perpetrators of such offences (Goodyer, 2013) (though the South African Civil Aviation Authority has banned the use of UAVs in their parks (Andrews, 2014)). 1.1.2 Disaster response and relief applications UAVs can be deployed (Apvrille et al., 2014) during or after disasters to organize disaster management operations, assist the population, reduce the number of victims and mitigate the economic consequences (Tanzi et al., 2014). Disaster scenarios are highly dynamic and authorities normally operate under imperfect information, which directly implies the importance of communication links (Tanzi et al., 2014), as well as the need for real-time changes in surveillance and data collection operations. UAVs can play a crucial role when existing infrastructure is compromised, malfunctioning or disconnected (Apvrille et al., 2014) or when the environment is deemed too dangerous for humans to operate in (Montufar et al., 2014) by delivering equipment, serving for geographical mapping or vehicular tracking (Giray, 2013). Significant resources have been allocated to develop supervisory control algorithms to assist a single operator and facilitate the remote controlling of multiple UAVs (Cummings et al., 2014). This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-2-W6-5-2017 | © Authors 2017. CC BY 4.0 License. 5 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2/W6, 2017 International Conference on Unmanned Aerial Vehicles in Geomatics, 4–7 September 2017, Bonn, Germany 2. SWARM BASED REAL-TIME DATA COLLECTION We address the scenarios where a number of UAVs are operating as a single functional unit (a swarm) to provide real-time data from their individual directed sensing equipment (such as onboard cameras). In this context, individual devices provide partial coverage which, when combined with the data from the other devices, offers complete coverage of a target object or area. The sensing capabilities of any equipment are bounded. Increasing the level of detail (e.g., the resolution of a camera) means reducing the area that is covered. If continuous coverage over an entire area is a hard constraint (as it is for certain search and rescue missions), then this can be achieved by handing over coverage over locations to other devices which currently operate under lower resolution requirem (...truncated)