Nonrecursive fixed-interval smoothing-based approach to attitude estimation

Abstract

We consider attitude estimation of objects equipped with gyroscopes in situations where fixed-interval smoothing is possible. A new nonrecursive attitude-estimation method based on Tikhonov regularization is proposed. The main benefit of the method is flexibility, which allows one to exploit observations of unknown accuracy involving an arbitrary number of time instances. Such observations are essential especially in sports applications, yet the existing attitude-estimation algorithms fail to provide a way to exploit them.