Advanced robots that can both ascend and navigate on land generally land on solid ground prior to attempting to change form and drive away. However, when the landing surface is uneven, these machines can occasionally become immobilized and fail to proceed. A team of engineers from Caltech has now created a real-world Transformer that possesses the “intelligence” to change form while airborne, enabling the drone-like robot to effortlessly roll away and begin its ground operations without interruption. The enhanced agility and durability of such machines could be especially beneficial for commercial delivery services and robotic explorers.
The new robot, called ATMO (aerially transforming morphobot), employs four thrusters for flight, but the protective shrouds transform into the system’s wheels in a different driving setup. The entire metamorphosis is dependent on a single motor that manipulates a central joint, raising ATMO’s thrusters into drone mode or lowering them into driving mode.
The scholars detail the robot and the advanced control system that operates it in a study recently released in the journal Communications Engineering.
“We devised and constructed a new robotic system inspired by the natural world—by how animals utilize their bodies in various manners to achieve different forms of movement,” states Ioannis Mandralis (MS ’22), a graduate student in aerospace at Caltech and primary author of the recent study. For instance, he explains that birds fly and then alter their body structure to decelerate and avoid impediments. “Having the capability to transform in the air opens up numerous possibilities for enhanced autonomy and resilience,” Mandralis comments.
However, midair transformation also introduces obstacles. Complicated aerodynamic forces come into play both due to the robot’s proximity to the ground and its shape alteration during the transformation process.
“Although it appears straightforward when observing a bird land and then run, this is a challenge that the aerospace sector has grappled with for more than 50 years,” says Mory Gharib (PhD ’83), the Hans W. Liepmann Professor of Aeronautics and Medical Engineering, director and Booth-Kresa Leadership Chair of Caltech’s Center for Autonomous Systems and Technologies (CAST), along with being the director of the Graduate Aerospace Laboratories at the California Institute of Technology (GALCIT). All flying vehicles encounter complex forces when near the ground. Consider a helicopter, for instance. As it descends for a landing, its thrusters direct a significant amount of air downwards. When that air impacts the ground, a portion is reflected back upward; if the helicopter descends too rapidly, it can be drawn into a vortex created by the rebounding air, resulting in a loss of lift.
In the case of ATMO, the level of complexity is heightened. The robot must not only manage complicated near-ground forces, but it also has four jets that continually adjust the extent to which they are directed toward each other, generating additional turbulence and instability.
To gain a deeper understanding of these intricate aerodynamic forces, the researchers conducted experiments in CAST’s drone laboratory. They utilized what are referred to as load cell experiments to examine how altering the robot’s configuration during landing influenced its thrust force. They also performed smoke visualization experiments to unveil the underlying phenomena that contribute to such changes in dynamics.
The researchers subsequently incorporated those insights into the algorithm that drives a new control system they established for ATMO. This system utilizes an advanced technique known as model predictive control, which operates by continuously forecasting how the system will behave in the near term and modifying its actions to stay on course.
“The control algorithm represents the principal innovation in this study,” Mandralis remarks. “Quadrotors typically employ specific controllers due to the configuration of their thrusters and their flight patterns. Here, we present a dynamic system that hasn’t been explored previously. As soon as the robot starts transforming, different dynamic couplings emerge—various forces interacting with one another. The control system must be capable of responding swiftly to all of that.”
Additional authors from Caltech on the paper “ATMO: An aerially transforming morphobot for dynamic ground-aerial transition” include Reza Nemovi, a design engineer in aerospace; and Richard M. Murray (BS ’85), the Thomas E. and Doris Everhart Professor of Control and Dynamical Systems and Bioengineering. Co-author Alireza Ramezani, an associate professor of electrical and computer engineering at Northeastern University, is presently visiting in aerospace at Caltech.