Otus Quadcopter 

Open development platform with motion capture, Raspberry Pi and the PX4® environment.

Teach, develop, and create in a few hours. Open and infinitely customizable. Precise position hold and navigation out of the box. See our documentation here.


Teaching and Research

High school and post-graduate research can benefit from reduced setup time and focus on teaching the concepts that matter.

Machine Learning

Open AI gym compatible (alpha). Run Tensorflow on-board in real time for offline or online machine learning.

Swarm and Vision

Supports 10+ Tracked Quad simultaneous and Raspberry Pi compatible with OpenCV.

Video shows an example of a step velocity input.

The Platform


  • Otus tracker and software
  • Pixhawk®, Dronecode and Dronekit environment
  • Raspberry Pi
  • 250 size quadcopter
  • Complete dynamical characterization


  • Open platform
  • Complete, detailed documentation
  • Amazing support

Pixhawk is a trademark of Lorenz Meier. Dronecode is a trademark of the Dronecode Project, Inc.

Download Datasheet

0.5 mm

Refresh Rate

250 Hz

Up to

13 quadcopters

Tracking Area

5 m x 5 m x 5 m

Simple, Fast, and Open Development

Step 1: Implement

Write simple Python or C++ code* so you can focus on the algorithm and quickly iterate. You have full control over the quadcopter programmatically for low and high-level functions.

*Or implement your algorithms in C++ in the PX4 firmware for high performance applications (above 100 hz)

Step 2: Simulate

Test your you code in a few seconds before you fly**.

**Or simulate using the powerful PX4 simulation tools with SITL or HITL, Gazebo or JMAVSim and more.

Step 3: Fly!

Fly almost anywhere with the precision motion capture tool that is the fastest to set up and the most compact on the market.


Example 1: Drone Following a Ground Robot.

Two Otus are used in this example. The Quadcopter is controlled manually, and the vehicle automatically targets the quadcopters.

Example 2: Neural Network Trained With Reinforcement Learning.

The Otus Quadcopter model, compatible with OpenAi Gym, was trained to target a location using the PPO reinforcement learning algorithm. The initial pose is random.

Code and Fly

Order Now