Estimating Motion Parameters of Head by Using Hybrid Extended Kalman Filter

Abstract

In this paper, we introduce an optical and inertial helmet tracker system that is required to estimate the motion parameters of the helmet. In this case, the motion parameters consist of the position, velocity, acceleration, attitude and angular velocity. This helmet tracker system consists of two infrared CCD sensors, several infrared LEDs on the helmet, an inertial measurement unit (IMU) and a computer for tracking. Our Helmet tracking system has a framework with a two-channel motion filter structure. The two channels, one for the optical measurement from the image sequence of CCD sensors and the other for the inertial measurement from the IMU, process independently with different sampling rates. With this hybrid system, we can overcome the failure in tracking a rapid motion and the accumulated error from the inertial sensors. Because of the nonlinearity in the state model, we implemented the system with the Extended Kalman Filter (EKF). The EKF has two channels for measurement that share a common prediction module. We implemented the real system and conducted the simulation with the real sensor data.