TY - CONF A1 - Häusler, F. A1 - Stührenberg, J. A1 - Neumann, Patrick P. ED - Sigalov, K. ED - Hagedorn, P. ED - Schönfelder, P. ED - Faltin, B. ED - Zentgraf, S. ED - Block, M. T1 - Gasquellenlokalisierung mit Drohnenschwärmen N2 - Gasquellenlokalisierungen (Gas Source Localization, GSL) tragen dazu bei, die Folgen von Industrieunfällen und Naturkatastrophen zu mildern. Während die GSL, wenn von Menschen durchgeführt, gefährlich und zeitaufwändig ist, können Schwärme von wendigen und kostengünstigen Nanodrohnen die Effizienz und Sicherheit der Suche erhöhen. Da die geringe Nutzlast von Nanodrohnen die Sensor- und Rechenressourcen einschränkt, werden Strategien zur Koordination des Roboterschwarms verwendet, die von biologischen Schwärmen, wie Kolonien sozialer Insekten, inspiriert sind. Die meisten Schwarm-GSL-Strategien verwenden das Maximum der Gaskonzentrationsverteilung zur Schätzung der Gasquellenposition. Experimente legen jedoch nahe, dass die Intermittenz der Gasverteilung vielversprechender ist. In diesem Beitrag wird eine neuartige GSL-Strategie für Schwärme vorgestellt, die auf Pheromonkommunikation und Intermittenz der Gasverteilung basiert. Die Agenten, d.h. die Nanodrohnen, emittieren Pheromonmarker in einer virtuellen Umgebung, wenn sie eine neue Gaswolke feststellen. Die Agenten werden durch virtuelle Kräfte gesteuert und nutzen abwechselnd das Wissen des Schwarms, indem sie dem Pheromongradienten folgen, oder erkunden den Suchraum, indem sie einen Zufallspunkt ansteuern. Zur Kollisionsvermeidung werden die Agenten durchgehend von anderen Agenten und Wänden abgestoßen. Die Strategie wurde auf drei Nanodrohnen implementiert und durch ein Experiment in einem Innenraum mit einer statischen Gasquelle validiert. Die Ergebnisse zeigen eine Verbesserung gegenüber maximabasierten Verfahren und geringe Lokalisierungsfehler in Windrichtung. T2 - 34. Forum Bauinformatik 2023 CY - Bochum, Germany DA - 06.08.2023 KW - Mobile Robotic Olfaction KW - Bouts KW - Mobile Sensorik KW - Nanodrohnen KW - Pheromonkommunikation KW - Schwarmrobotik PY - 2023 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-594320 DO - https://doi.org/10.13154/294-10088 SP - 168 EP - 175 AN - OPUS4-59432 LA - deu AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Neumann, Patrick P. A1 - Häusler, F. T1 - Bout-based Gas Source Localization using Aerial Robot Swarms N2 - Gas source localization (GSL) helps mitigate the impact of industrial accidents and natural disasters. While GSL may be dangerous and time-consuming when performed by humans, swarms of agile and inexpensive nano aerial robots may increase the safety and efficiency of gas source localizations. Since the small payloads of nano aerial robots limit the sensing and computing resources, strategies adapted from biological swarms, such as colonies of social insects, are used to coordinate robot swarms. Most swarm GSL strategies are based on the assumption that the maxima of gas concentrations are sufficiently close to the gas sources. However, prior studies have indicated that the occurrence of “bouts”, a metric for the intermittency of gas distributions, may advantageously be used as a more accurate gas source proximity indicator. This paper presents a swarm GSL strategy employing bouts as source proximity indicators and a bio-inspired pheromone system for communication. Nano aerial robots, deployed in this study, act as agents and emit pheromone markers in an artificial environment upon detecting bouts. Leveraging the concept of artificial potential fields, the agents switch between exploiting the knowledge of the swarm by following pheromone gradients and exploring the search space by targeting a random point. The agents are repelled by each other and by walls to avoid collisions. The swarm GSL strategy is implemented into three nano aerial robots and validated in a real-world experiment in an indoor environment with a single gas source. The results indicate that the the swarm GSL strategy presented in this paper is capable of GSL in indoor environments and that the intermittency of gas distributions is a better source proximity indicator than the mean concentration. T2 - IEEE Sensors 2023 CY - Vienna, Austria DA - 29.10.2023 KW - Mobile robotic olfaction KW - Nano aerial robot KW - Gas source localization KW - Bouts KW - Distributed robotics PY - 2023 AN - OPUS4-58779 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Häusler, F. A1 - Stührenberg, J. A1 - Smarsly, K. A1 - Neumann, Patrick P. T1 - Bout-based Gas Source Localization using Aerial Robot Swarms N2 - Gas source localization (GSL) helps mitigate the impact of industrial accidents and natural disasters. While GSL may be dangerous and time-consuming when performed by humans, swarms of agile and inexpensive nano aerial robots may increase the safety and efficiency of gas source localizations. Since the small payloads of nano aerial robots limit the sensing and computing resources, strategies adapted from biological swarms, such as colonies of social insects, are used to coordinate robot swarms. Most swarm GSL strategies are based on the assumption that the maxima of gas concentrations are sufficiently close to the gas sources. However, prior studies have indicated that the occurrence of “bouts”, a metric for the intermittency of gas distributions, may advantageously be used as a more accurate gas source proximity indicator. This paper presents a swarm GSL strategy employing bouts as source proximity indicators and a bio-inspired pheromone system for communication. Nano aerial robots, deployed in this study, act as agents and emit pheromone markers in an artificial environment upon detecting bouts. Leveraging the concept of artificial potential fields, the agents switch between exploiting the knowledge of the swarm by following pheromone gradients and exploring the search space by targeting a random point. The agents are repelled by each other and by walls to avoid collisions. The swarm GSL strategy is implemented into three nano aerial robots and validated in a real-world experiment in an indoor environment with a single gas source. The results indicate that the the swarm GSL strategy presented in this paper is capable of GSL in indoor environments and that the intermittency of gas distributions is a better source proximity indicator than the mean concentration. T2 - IEEE Sensors 2023 CY - Vienna, Austria DA - 29.10.2023 KW - Mobile robotic olfaction KW - Nano aerial robot KW - Gas source localization KW - Bouts KW - Distributed robotics PY - 2023 SN - 979-8-3503-0387-2 DO - https://doi.org/10.1109/sensors56945.2023.10325109 SP - 1 EP - 4 PB - IEEE AN - OPUS4-58778 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -