Approximation of discrete data by \(G^ 1\) arc splines (Q1195289)

The authors develop algorithms for fitting a \(G^ 1\) arc spline to data so that the data points are within a specified distance of the arc spline. The \(G^ 1\) arc spline is made of biarcs that are joined in a \(G^ 1\) manner. These methods adaptively find the number of biarcs to use for the approximation. These methods do not use least squares and the solution of a linear system of equations but use standard formulae to estimate tangents from discrete data and standard algorithms to partition the discrete points for the approximation by biarcs.