Next:Computation of the Frenet Up:The system of first Previous:Euler's method
 
 

2.2 Bulirsch-Stoer method

 

    This method is also based on stepwise extrapolation defined by an initial stepsize   z, however, one single step usually consists of many substeps of the so called modified midpoint method which are extrapolated to zero stepsize using a rational function (Richardson extrapolation). Both the consecutive stepsizes and the number of substeps are determined as functions of a predefined accuracy limit. Its value was set to 10-6 m in the computations.