This weekend in London, Imperial College showcased their state of the art facilities, labs, lectures and work from the students at one large festival. The main focus this year was on robotics and biology and lecturers Professor Holger Krapp and Dr Mirko Kovac combined the two perfectly. These two combined their expertise and interests in separate fields to produce better engineered flying robots and also understand the natural engineering that occurs in biology of flying animals. The main concept was stability, when you look at how various species of fly take off and approach certain surfaces they all act in unique ways. For example, there was one species of fly that somersaults completely whilst taking off; whilst we don’t plan on making flying drones complete unnecessary stunts like these, the small manoeuvres of the fly’s wings provide a useful insight into efficient take-off and landing. The conventional drone structure that we see in commercial drones consists of a few propellers in a quadcopter style arrangement, this is for stability but doesn’t have much resemblance to nature’s design for flying animals. When looking for inspiration at jet fighters and flies they are both very streamlined and agile, however the jet fighter isn’t as stable. When looking a slow motion recording of a hoverfly, you can see that it keeps its head completely level even when the body is rotating with angular velocities of 5000 degrees per second. Thus, if we are going to adapt some of nature’s engineering into our drone designs we may be looking at smaller, streamlined drones hovering through the air like an insect.
Throughout the robotics exhibition there were many projects focused on the way in which flies can help us understand how to develop better technology in UAVs. Behind the commercial hype of drones there are many with specific uses: drones that could possibly create structures; drones that could detect where they are needed, for example detecting a forest fire and delivering aid before it is instructed to do so; drones that could even possibly deliver parcels, a concept not so far-fetched with projects such as Amazon Prime Air being developed. The key to creating these efficient drones is reducing the computing time. A fly uses its compound eyes to decipher how to move its wings and create an image of its surroundings. Whilst many people may think the compound eyes create a high definition image which provides more information to the fly, the images of the fly’s surroundings are actually of very poor quality, partly down to the fact the flies want to use this information quickly. The fly’s sight is much more focused on field of view and depth, viewing an optic flow image almost in some kind of vector field where the fly’s actions can be understood by the distorted areas due to motion. As an animal walks it seems their surroundings are moving around them, in a drone where multiple still images are pieced together this takes a lot of computation time, if a drone were to capture images like a fly focussing on depth and direction, the computer in the drone would use less energy and less time.
The fast development of UAVs in recent years has fascinated us all and the increasing list of their uses is impressive, but the main limitations to all of them seem to be their energy consumption and storage. These robots are small and therefore can’t carry much energy, neither would solar energy be enough alone. Drones designed to deliver aid to remote areas are an excellent proposition but if the drone only has an hour of battery the whole idea is redundant. The solution isn’t only looking at alternative power sources, but reducing the power consumption. By looking to nature we can simplify technology to overcome this barrier, because until drones are redesigned, innovation in this field can’t continue.