Pigeon Lab

Today, to deliver a 1 kg pizza, a 70 kg human rides a 110 kg bike—resulting in an overhead of 180 kg just to move a 1 kg payload. Our UAV weighs only 2 kg to deliver that exact same payload. By drastically reducing this physical overhead and eliminating the need for a human rider, our delivery cost drops from an average of ₹60 to just ₹6.

Roads are rarely a straight line from source to destination, and urban traffic only slows things down further. Our UAVs fly directly to their destinations at speeds of up to 100 km/h, bypassing congested streets and traffic signals to achieve true 10-minute delivery.

Helicopters can take off and land anywhere, but they can't travel far. Airplanes can cover great distances, but they require a runway. Our Hybrid Vertical Take-Off and Landing (VTOL) aircraft—the HeliPlane for short—can take off and land on rooftops while maintaining long-range flight capabilities. This increases a store's delivery radius to 20 km, allowing businesses to centralize their operations in a single location rather than opening branches all over the city.

To fly from one point to another, or even just to take off from the ground, the UAV continuously calculates its motor speeds to maintain its travel path. It does so reliably—even in strong winds, or when compensating for motor degradation and noisy sensor data. We rigorously test our control algorithms in simulated environments to ensure peak performance under all of these conditions.

To fly safely, our UAV perceives its surroundings in 3D. It does this the same way most animals do: with a pair of cameras. Just as closing one eye shifts the horizontal perspective of closer objects more than farther ones, we exploit this parallax effect to calculate exactly how far away an obstacle is, allowing the drone to navigate safely through free space.

Using this 3D perception data, the UAV plans its flight path to the destination while actively avoiding buildings and obstacles. It does this by representing the world as a dynamic grid of occupied and free space, instantly calculating the safest and most efficient route to its target.

High-rise buildings in urban areas can frequently block GPS signals. To solve this, our UAV navigates using its own internal memory map of landmarks, much like you do when commuting to the office. Technically, the UAV utilizes Simultaneous Localization and Mapping (SLAM). By recording distinct spatial points with its cameras, the drone can localize itself whenever it recognizes those landmarks again. This enables the UAV to navigate flawlessly without relying on external communication.

akshit@pigeonlab.in

+91-9581830311

linkedin.com/in/akshitj1