Drone miniaturization – a journey toward smart dust
Microdrones and smart dust
Smart Dust is defined as a swarm of free-flowing micro-electromechanical systems. A subcategory of smart dust is the dust that is self-propelled. In this case, the user has control over where the individual particles/motes go, making them easier to recover for clean up or reuse. Basically, this is a swarm of airborne micro-drones! What size factor do we need to reach for a swarm of micro-drones to qualify it as smart dust? When will a farmer be able to control a swarm of intelligent drones that could pollinize all the flowers of his vertical farming crops in a matter of minutes before returning to the palm of his hands? Will engineers be able to send smart dust to inspect the nooks and crannies inside an aircraft turbine to feed a digital twin and precisely estimate the lifetime increasing cost efficiency and safety? How far away are we from achieving this in terms of miniaturization?
Increasing regulation but continued innovation
There are increasing regulations related to drones, such as mandatory registration, and no fly zones; it is then becoming more difficult for drone enthusiasts to fly their machines. However, it would be a mistake to interpret this as the end of the drone era.
Drone use for industry applications such as pipeline or powerline inspections is today common. Even hospitality applications, such as drone assistants, are being tested out. In the consumer market, drone innovations keep coming, such as the Airselfie 2 marketed as the ultimate replacement of the selfie stick… and self-flying drones such as the Skydio R1 with advanced tracking and navigation systems. In terms of miniaturization, for the everyday consumer, there are remote controlled drones such as the JJRC H36 that fit in the palm of your hand for less than 15 $US! There is also a whole new world of competitive drone racing, which is emerging. Check out the regulated DRL (Drone racing league) with the sponsor Allianz, BMW, Swatch, and the U.S. Air Force. Lockheed Martin is partnering with the DRL and Nvidia for the AlphaPilot Challenge focusing on advance artificial intelligence and autonomous flight of racing drones. The Micro Air Vehicle Laboratory (MAVLab) of TU Delft has created a 72g drone they claim is the smallest autonomous racing drone in the world. They demonstrated it can fly loops through and obstacle course at an average of 2m/s!
Flight at small scales
As we move toward autonomous flight, at the insect scale, there are several hurdles. Airflow tends to be turbulent at these scales with disturbances often matching the flight speed of the vehicle. Also we do not yet fully comprehend how insects fly in low Reynolds aerodynamic mode. Typical aircraft flight is characterized by high Reynolds numbers (the viscosity of the fluid is negligible when compared to the inertial forces) when the Reynolds number gets very very small, viscosity is the dominant factor and the inertial component is negligible (swimming bacteria). In the intermediate range that characterizes insect flight viscosity and inertia both play an important role making this regime very difficult to model using the full navier stokes equation.
Even among insects there seems to be a significant difference between larger scale insect flight such as bees or dragonflies (2 to 5 cm) and tiny insects’ flight such as thrips and fairyflies (300 to 1000 microns).
Thermal combustion engines cannot be used because as they shrink, the relation between volume inside and envelop deteriorates, making thermal losses through the envelop increase and the efficiency go down. Also engineering a micro thermal combustion engine is a challenge because there are many moving parts. Currently piezoelectrics are the way to go at that scale.
This leaves batteries and their low energy densities as the only option for now. Energy density of batteries is a real problem and has only been improving by around 3% per year (far from the advance rates in microelectronics and moor’s law). A major innovation in batteries would be needed for micro aerial vehicles – there are high hopes that quantum computing applied to quantum chemistry might help us achieve this! The sensing and control algorithms that need to be embarked can currently only be run on relatively heavy hardware and need to be refined. However nature shows that this is possible in the brains of insects with only a few hundred neurons.
A team from the MIT has recently demonstrated untethered flight of an insect-sized flapping-wing microscale aerial vehicle side stepping the battery density problem by using an array of solar panels and artificial bright light. Also the vehicle in question cannot really be qualified as autonomous or even as particularly stable since it stays airborne only for less than a second before veering off to one side or the other. I believe all these developments, and the pace of innovation tend to show that an insect sized swarm of micro-drones should be considered as a form of self-propelled aerial smart dust, and its presence in our life is imminent.