Conventionally, the filtering technique for attitude estimation is performed using gyros or attitude dynamics

models. In order to extend the application range of an attitude filter, this paper proposes a quaternionbased

filtering framework for gyroless attitude estimation without an attitude dynamics model. The attitude

estimation system is established based on a quaternion kinematic equation and vector observation models.

The angular velocity in the system is determined through observation vectors from attitude sensors and the

statistical properties of the angular velocity error are analysed. A Kalman filter is applied to estimate the

attitude error such that the effect from the angular velocity error is compensated with its statistical properties

at each sampling moment. A numerical simulation example is presented to illustrate the performance of the

proposed algorithm.