Unmanned aerial vehicles, like quadcopters, are easily the most versatile types of robots when it comes to the places that they can go. Difficult terrain is just not a problem when it can simply be flown over. However, that versatility comes at a steep cost — namely, heavy energy expenditures. The constant spinning of the rotors drains batteries very quickly, which greatly limits the range of these vehicles.
Wheeled vehicles, on the other hand, use far less energy for locomotion, which gives them a much greater range. But they do lose the versatility of an aerial vehicle. In an effort to get the best of both worlds, researchers have been experimenting with vehicles that can transition between rolling and flying states, conserving energy when possible, then taking to the air when the going gets too tough.
A mid-air transformation allows for wheeled locomotion (📷: Caltech)
The process of transitioning between flight and wheeled locomotion is not a completely solved problem, however. In some cases, the vehicles first land, then transform. But when the terrain is rough, these vehicles sometimes get stuck. Other vehicles attempt to transform in mid-air, but in these instances, complex aerodynamic forces come into play that frequently lead to instability or crashes.
A team led by researchers at Caltech has just proposed a new solution to this problem for vehicles that fall into the latter category. Called the Aerially Transforming Morphobot (ATMO), their transforming robot changes shape in mid-air, but it does so in an intelligent way that prevents the instability that plagued vehicles of the past. This allows it to set down as a wheeled robot that can smoothly drive away without missing a beat.
ATMO has a unique design in which the hardware serves dual purposes. Its four thrusters are housed within protective shrouds that also function as wheels when the robot is on the ground. A single motor powers a central joint that shifts the robot between flight and drive modes, minimizing the weight and complexity typically associated with multimodal vehicles.
To overcome the aerodynamic instability that mid-air transformation introduces — especially near the ground, where turbulence and the so-called “ground effect” complicate flight — the researchers conducted extensive lab tests. These included load cell experiments to measure thrust forces during transformation, and smoke visualization experiments to map the airflow around the robot.
The insights from these tests fed directly into the development of ATMO’s control system, which uses a technique called model predictive control. This algorithm continuously forecasts the robot’s behavior and adjusts motor commands in real time, allowing it to stay stable and controlled throughout the transition.
The robot’s ability to transition between flight and ground movement make it a good candidate for a wide variety of applications where extra range is needed, from commercial package delivery to planetary exploration. By avoiding redundant hardware and enabling operation in complex environments, ATMO could move the entire field forward.