The premise
Drones are expensive to run. Hovering costs energy continuously just to stay up, and a battery that could drive a vehicle for an hour will fly one for a few minutes.
So: where the ground is drivable, don’t fly. Fly to the site, land, and become a ground vehicle.
That argument was made with numbers rather than asserted — a three-state LiPo decay model (exponential, linear, inverse) simulated in MATLAB, comparing endurance on BLDC flight motors against geared DC drive motors. The gap is not small, and it is the entire justification for the machine.
What it does
In the air it is a quadrotor. On the ground the rotor arms fold down and the shrouded prop-guards become the wheels — the same appendages, repurposed, driven by geared DC motors through a gear train. Two 60 kg·cm servos handle the transition.
Then it reaches. A 4-axis arm actuated by shape memory alloy — three SMA springs in parallel, 5 V, 2.17 A — with a gripper, steered by hand gestures through a vision pipeline. SMA was chosen because a disaster arm has to be light, and conventional actuators at that scale are mostly mass.
The mission it was designed around: fly over a collapse scanning for signs of life, land, drive through rubble and gaps too tight or unstable to fly in, and retrieve small objects — a water bottle, a survival kit — for someone trapped.
Two prototypes, and the number that changed
Both were analysed the same way: FEA static stress on the arms, wheels and frame, and CFD using the Reynolds-Averaged Navier–Stokes equations at 5 m/s.
The second prototype streamlined the wheel spokes to reduce drag, and the result is the clearest engineering win in the project:
| prototype 1 | prototype 2 | |
|---|---|---|
| coefficient of drag | 0.015 | 0.0082 |
| traction, Z axis | 0.244 N/m² | 0.133 N/m² |
Drag nearly halved — from spoke geometry alone. On a vehicle whose whole argument is energy efficiency, that is the argument being won.
Arms and wheels were printed in Hyper PLA with a carbon-fibre frame. The FEA was run on ABS as a substitute, since Hyper PLA wasn’t in the material library — a compromise worth stating rather than hiding.
Credit where it’s due
The multi-modal morphing concept follows published work on appendage repurposing for locomotion plasticity (Sihite et al., Nature Communications, 2023). This is an implementation of that idea with an SMA manipulator added, not a claim to have invented it.
Built with Soham Mondal and Tejal Uplenchwar under Prof. I.A. Palani, Mechatronics and Instrumentation Lab, IIT Indore.