Inspiration from Ants Dutch Scientists Develop Navigation System for Micro Drones
TapTechNews July 20th news, Micro-robots, micro-drones and other products have a large number of potential uses, such as searching for survivors in disaster areas, conducting reconnaissance in dangerous environments, pollinating crops, etc. They need to fly autonomously to a specific area and then return to the base. However, due to their limited on-board capacity, they cannot carry heavy and power-consuming facilities such as sensors and processors, so the navigation ability of these micro-intelligent devices is severely limited.
According to the comprehensive report of NEWATLAS and CCTV News today, scientists at Delft University of Technology (TUDelft) in the Netherlands studied ants and other foraging insects and were inspired by the visual navigation ability of such insects.
In previous research, ants and other insects can combine the odometer measurement of their own movement with the visually guided behavior based on snapshots, thereby navigating for themselves. During the journey, ants will take snapshots of the surrounding environment from time to time using their own visual system, which is used as the basis for finding these positions again.
And after arriving at the corresponding position of the previously taken snapshot, ants will compare the current visual perception with the snapshot again, and move to the place where the difference between the seen environment and the snapshot is the smallest, which can make it navigate back to the position where the snapshot was taken before. And so on, until it reaches the nest.
Under the leadership of professors Tom van Dijk and Guid de Croon, the TUDelft team applied this principle to a 56-gram micro quadcopter drone and equipped it with an omnidirectional camera. Of course, the aerial drone does not walk like an ant, so the number of steps cannot be calculated.
To measure distance, our drone uses a method similar to that of bees, calculating movement through the optical flow, de Croon explained. For this, our robot has a small downward-facing camera that tracks the speed at which objects pass through the visual field.
On the return journey, once the drone determines that it has departed from a recorded snapshot according to the recorded distance/direction, it will compare the current camera image with the next recorded snapshot. Considering that the drone will inevitably have some drift during the return journey, it will correct the heading until the seen scene almost completely matches the snapshot.
In indoor environments, navigating in this way, the drone is able to autonomously return to the base along a tortuous 100-meter obstacle route, using only 1.16 KB of memory.
The relevant papers have been published in the American journal Science Robotics, and TapTechNews attaches the link: Click here to go.