STochastic Algorithm for Robust Guidance Analysis and Trajectory Estimation

We are approaching a new space age, where safety and cost reductions are key factors to achieve a more democratic space access.

A crucial component of this new era will be represented by autonomous systems able to safely complete their mission despite a harsh and vastly uncertain environment, (for instance the atmosphere of Mars or the gravitational field of an asteroid).

A disruptive example of autonomous systems is given by SpaceX’s rockets, which are safely reused, thus drastically easing human access to space.The current optimal trajectory generation process uses deterministic models. However, this approach is valid as long as the real system does not excessively deviate from its abstract model.

When this happens, a (sometimes critical) loss of performance occurs, due to the uncertainties’ effects playing a major role in real-world scenarios.STARGATE has the groundbreaking goal to change how trajectories for autonomous systems are computed by including uncertainties in the design through their stochastic properties rather than simply assess their effect with a-posteriori validation.

The proposed methodology will overcome the so-called curse of dimensionality, a so far insurmountable technological limit for these applications, by using a novel, scalable approach coming from the interplay of two new theories, Unified Unscented Transform and Kalman-Pontryagin Sequential Consensus Optimization.

I will forge the theoretical formulation and a numerical framework accompanied by realistic applications (e.g., Six-DoF rocket landing, asteroid descent, and spacecraft rendezvous) to demonstrate the versatility of this brand-new methodology. Finally, I will move to a real-testing stage by using world-class testbenches available at DLR.

With my team made of 2 PhDs and a Postdoc I plan to pursue in the next 5-years the vision behind STARGATE: make a new, scalable and certifiable methodology to design robust optimal trajectories for autonomous systems a technological reality.