PDF version of this document
Since some of the gathered statistics indicate that the program achieved what it had been set to achieve, there are intentions of applying the same concepts to 2-D and data in the foreseeable future. The principles remain unchanged and the only required extension is that of CPS to a higher-dimensional space - something which has been developed already.
This extension step is expected to be trivial as CPS warps in 2-D have already been dealt with by Marsland and necessary 2-D data is available in the Division. Synthetic 2-D data can be generated as well if necessary and code exists for doing so (see Figure cap:2-D-Synthetic-data). Once generation of data becomes possible and shapes for which the estimated solution is known can be created, e.g. triangles in 2-D space, then well-controlled 2-D tests can be performed under AART.
The only envisioned hindrances will then be the speed of execution and the diligent selection of knot-points for transformation. Since there is yet some lacking understand of the problem in 1-D, time is needed to improve the algorithm. Its requirements need to have a more responsive time-span.