in [12] that many properties of polynomial spatial PH curves can be transferred to the whole
family of rational curves with the same tangent indicatrix. This family was first investigated
in [11] where the envelope approach to planar rational PH curves of Pottmann and Farouki
[6] was generalized to rational PH space curves. While being rather straightforward and
comprehensive, it seems that degree of the denominator polynomial and curve shape are
difficult to control. The dual approach was continued in series of papers [19–21] where the
authors solve interpolation problems with rational spatial PH curves of low class. In [9] a
special form of the rational hodograph is used to construct planar rational PH curves with
rational arc-length function.
In the recent article [18], we have (together with co-authors) suggested a novel computa-
tion method for rational spatial PH curves. Motivated by the theory of motion polynomials
in the dual quaternion model of space kinematics, c.f. [13, 15, 22] we identified rational
PH curves with so called framing rational motions. Using suitable factorization we sep-
arated the spherical and translational part of the motion. The spherical component has
been called Euler-Rodrigues motion or, equivalently referring to the images of a suitable
orthogonal tripod, Euler-Rodrigues frame [1, 19]. Selecting appropriate parameters (the
spherical motion component, represented by a quaternion polynomial A ∈ H[t]and the
PH curve’s denominator polynomial α) the remaining coefficients can then be computed
by solving a modestly sized and well-structured system of linear equations.
Here, we continue this approach. While [18] was mostly concerned with the existence
of truly rational (non-polynomial) solutions, we now investigate the complete structure
of the solution spaces. While the formalism of motion polynomials yields a very natural
and compact way of understanding this problem, it is not completely necessary and we
abandoned it in this paper for two reasons. The first reason is to make our results better
understandable for a wider audience. The second one is to make it obvious that our ap-
proach generalizes to more general cases of the tangent fields and also to higher dimensions.
Our central result is Theorem 3.12 that, together with Theorems 3.6, 3.10, and 3.13, allows
us to construct canonical bases for solution spaces and also results in a rather surprising
partial fraction decomposition for solution curves (Corollary 3.16).
We proceed with introducing some necessary basic information in Section 2. Our main
results are presented in Section 3, at first for PH curves with a single root in the denom-
inator (Section 3.1) and then for general denominators (Section 3.2). In order to make
Section 3 as clear as possible also several examples are included there. The corresponding
technical proofs are collected in Section 4 in order to allow for an uninterrupted flow of
reading in Section 3. Our main tool of proof is a careful but rather technical study of an
underlying system of linear equations in the special case of single root denominators. In
concluding Section 5 we provide a more general understanding of our results as well as
some possible directions for the future research.
2. Preliminaries
In this section we will state the problem which we propose to solve. We will also
place our approach in the context of existing constructions, in particular in the context
2