Researchers from Delft University of Technology demonstrated a swarm of tiny drones equipped with optical sensors that can explore unknown environments completely by themselves. The concept, inspired by insect swarms, relies on numerous inexpensive robots with relatively limited sensing and computational techniques, which together are able to explore large areas with greater speed and efficiency than a single larger drone. “The biggest challenge in achieving swarm exploration lies at the level of the individual intelligence of the drones,” said Kimberly McGuire, a Ph.D. student who worked on the project. The researchers first tackled basic flight capabilities before devising a method to ensure that the drones didn’t run into each other. For that, the researchers had each drone carry a wireless communications chip, and then made use of the signal strength between the chips to avoid collision, a method that requires very few computations, McGuire said. “They are equipped with a minimal set of sensors, since they are so very small,” Guide de Croon, principal investigator of the project, told Photonics Media. The drones have an inertial measurement unit consisting of accelerometers and gyros, an optical flow computer mouse sensor, and a tiny downward pointing laser — “This is to measure the drone’s horizontal velocity,” de Croon said — and four other lasers that measure a point distance to the four sides of the drone. “The lasers are set up as a cross, pointing forward, backward, left, and right,” de Croon told Photonics Media. “We showed that with these sensors the drone were able to explore an unknown environment all by themselves.” For the project, the drones were equipped with cameras and deployed in an indoor office environment measuring roughly 40 × 12 m to find two dummies representing victims in a disaster scenario. “In the experiment, the ‘rescue’ workers could look at these images to see if and where there were ‘victims’ in the environment,” de Croon said. Within six minutes, a swarm of six drones was able to explore approximately 80% of the building, a feat which would be impossible for one of the drones on its own. Further, the swarm function proved to be a useful redundancy. One drone found a victim but was unable to bring back images due to a hardware failure. Luckily, another drone captured the victim on camera as well. “The main idea underlying the new navigation method is to reduce our navigation expectations to the extreme: We only require the robots to be able to navigate back to the base station,” de Croon said. “The swarm of robots first spreads out into the environment by having each robot follow a different preferred direction. After exploring, the robots return to a wireless beacon located at the base station.” The navigation method, McGuire said, is a novel type of bug algorithm. Insects do not make highly detailed maps, but instead retain landmarks and behaviorally relevant places such as food sources and their nest. “In principle, detailed maps are very convenient because they allow a robot to navigate from any point in the map to any other point, along an optimal path. However, the costs of making such a map on tiny robots is prohibitive,” McGuire said. “The proposed bug algorithm leads to less efficient paths, but has the merit that it can even be implemented on tiny robots.” In the future, de Croon told Photonics Media, the drones could detect victims with onboard processing and stay and beep loudly, or could signal other drones in the swarm of a discovery and form a chain from the victim to the base station that could be followed by the rescue worker. The research was published in Science Robotics (www.doi.org/10.1126/scirobotics.aaw9710).