Developments in the planning field have enabled automated planners to reason with continuous change. Several planners exist today that can handle linear continuous change, however, scalabiliity remains a challenge for planners capable of reasoning with non-linear domains. This paper applies a novel approach to reasoning with non-linear domains to Aerospace related problems. Using linear over and under estimators to bound the problem, then using well proved scalable linear planners to find a valid plan. The usefulness of the linearizion approach is shown by preprocessing the nonlinear domain to produce a bounding linear domain then solving the planning problem. The advantages and disadvantages of the method are demonstrated on a rocket vertical landing domain, and a solar rover domain.
|Title of host publication
|58th Israel Annual Conference on Aerospace Sciences
|Published - 6 Mar 2018