Novel Structure-Preserving Algorithms for Accurate Rocket Trajectory Propagation

The Department of Defense uses large-scale high-resolution federated simulations to propagate rocket vehicle trajectories. Runge-Kutta methods have served as a de-facto standard while conducting such simulations. However, there are several challenges while using Runge-Kutta methods for this task. Firstly, there should be exact time-step matching between federates, otherwise the states have to be interpolated between time steps, further introducing errors. Secondly, Runge-Kutta methods requires extensive computational throughput which in-turn increases the data output size. Thirdly, rocket vehicles follow different dynamics at different stages of its trajectory. Runge-Kutta integration schemes incur large integration errors at instances when change in dynamics occur. The proposed work develops a new class of structure preserving algorithms which inherently addresses all these limitations. Constants of motion are preserved throughout the rocket trajectory ensuring better error control at integration points and allowing larger time-steps, thus reducing output data. Moreover, a structure preserving interpolation method is proposed ensuring that the in-between states follow laws of physics obeyed by the underlying dynamical system. Phase I will demonstrate the feasibility of the proposed algorithms. In Phase II, a prototype will be developed to run multiple federated simulations.Approved for Public Release | 18-MDA-9522 (23 Feb 18)