Sensor Fusion for Object Tracking



Tracking in modern commercial VR systems is based on the principle of sensor fusion, where measurements from multiple independent sensors are combined to estimate the position and orientation of tracked objects with better quality than what can be achieved by using those same sensors in isolation.

This video shows a simulation of a moving and rotating object in two dimensions, tracked by an external absolute measurement system and a relative measurement system integrated into the tracked object. Measurements from these two systems are combined using a non-linear extension of the Kalman filter, yielding a result with low noise, low update latency, and no drift.

Related videos:
Pure IMU-based Positional Tracking is a No-go: https://www.youtube.com/watch?v=_q_8d0E3tDk
Optical 3D Pose Estimation of Oculus Rift DK2: https://www.youtube.com/watch?v=X4G6_zt1qKY
Lighthouse Tracking Examined – Headset at Rest: https://www.youtube.com/watch?v=Uzv2H3PDPDg
Lighthouse Tracking Examined – Headset in Motion: https://www.youtube.com/watch?v=XwxwMruEE7Y
Lighthouse Tracking Examined – Controller in Motion: https://www.youtube.com/watch?v=A75uKqA67FI
Playstation Move Tracking Test: https://www.youtube.com/watch?v=0J5LaWykiIU

More information:
https://en.wikipedia.org/wiki/Kalman_filter

source

Leave a Reply


*