Prototype drones capable of navigating dangerous and unpredictable environments without crashing could prove useful for search-and-rescue teams.
at the Norwegian University of Science and Technology and his colleagues have developed a flying drone that aims to avoid crashes altogether. The robot, called RMF-Owl, made its debut while winning a competition hosted by the US Defense Advanced Research Projects Agency, in which it had to navigate an underground mine and perform rescue-related tasks.
De Petris and his team designed the drone to focus on avoiding crashes, while it mimics rescue operations in cramped, complex environments. The RMF-Owl, which dodges collisions via laser mapping and onboard algorithms, can also work with other flying and walking robots.
Advertisement
鈥淚f the robot crashes in the middle of a passage, which is really [narrow], then all the other robots that come afterwards will be blocked and cannot proceed in their navigation, so we wanted to avoid that,鈥� says De Petris.
The flying drone, which weighs around 1.5 kilograms, has a carbon-foam frame in case the collision-avoidance system fails. However, this didn鈥檛 occur in its tests.
鈥淚t鈥檚 an impressive capability,鈥� says at the Swiss Federal Institute of Technology in Lausanne. 鈥淭hey travel quite a long distance inside of a mine without having any collisions and by generating their own map and moving within their map.鈥�
at Montreal Polytechnic in Canada and his colleagues took a different approach towards avoiding drone crashes. Inspired by the flexible exoskeletons of insects, they designed the CogniFly, a small, autonomous drone weighing around 250 grams that can survive repeated high-impact collisions.
鈥淲e were doing experiments and mostly we were crashing [the drone] against the floor, so we needed something to absorb the crash landings,鈥� says Azambuja.
The team has designed a shell that absorbs landing impacts like a cricket鈥檚 legs, with the drone being built from 3D-printed parts and having flexible joints. During normal drone experiments, propellers often break from collisions and have to be replaced. The CogniFly鈥檚 shock absorbers meant the team only required two sets of replacement propellers during the drone鈥檚 tests. Azambuja says the durability of the robot could be useful for unpredictable scenarios, such as during cave exploration.
鈥淭he trade-off between the two robots is like hardware improvement versus software improvement,鈥� says Zufferey. 鈥淚t鈥檚 nice to see this duality here between these two papers, it鈥檚 two approaches to solve a similar problem.鈥�
References: &
