We're building autonomous aerial harvesters for Kashmir's mountain orchards — engineered with physics-informed AI to stay stable in terrain where every other drone fails.
Every harvest season, families send their sons up 40-foot walnut trees with bamboo poles. Every season, some of them don't come down safely — and the fruit that does get picked often arrives bruised, dropping straight out of export grade.
Most drones fight the mountain wind and lose. Ours holds onto the branch first — then does the delicate part.
A mechanical jaw locks onto a structural branch, so the drone borrows the tree's stability instead of fighting the wind on rotor power alone.
Onboard vision identifies ripe fruit and reads its exact position, while a physics-informed flight model anticipates the next gust before it hits.
A soft robotic arm grips with just enough pressure for the fruit — gentle for apple and cherry, firmer for walnut — and releases it bruise-free.
I grew up watching harvest season in Kashmir up close — the pride in a good crop, and the quiet risk every family accepts to bring it in. That risk shouldn't be the cost of doing the work.
My background sits right at the intersection this problem needs — mechanical engineering from NIT Srinagar, and current research applying AI to heavy machinery at IIT Kharagpur. Robus Aerospace is where that research meets the orchard floor.