Team Members: Eden Dolev, Elisha Gur

Supervisors / Mentors: Josh Karon

 

As part of a collaborative project with SixDof Space, we developed an automated calibration system for high-precision motion tracking using infrared (IR) sensors. Accurate calibration is essential for determining the precise position and orientation (6 degrees of freedom) of each sensor in space. Traditionally, this process involves manual measurements, which are slow and error-prone.

Our system replaces manual calibration with a fast, reliable, and fully automated alternative. It uses a handheld calibration wand with known geometry, visible to all sensors, as the reference object. An initial estimate of each sensor's pose is computed using a geometric solver (P3P), and then refined through graph-based optimization with the g2o framework. This yields a consistent and accurate sensor configuration without any manual input.

The calibration process completes in under 10 minutes, achieves sub-centimeter accuracy, and requires only that the user wave the wand through the tracking area. The system is designed for applications in robotics, augmented reality, and medical navigation — wherever precise real-time tracking is critical.