Astrograd is a Python astrodynamics library that combines automatic differentiation, hardware-accelerated parallel execution and embedded code generation with models specific to Guidance, Navigation and Control applications.
The library is built around a single idea: assemble a differentiable simulator once from reusable modules, then run it unmodified across many numerical computing backends.
Simulators built with astrograd can be coupled with deep-learning components and executed with massive parallelism on machine-learning frameworks such as JAX, PyTorch or TensorFlow.
The same models work directly with CasADi for optimal control and estimation, or can be compiled to embedded code for flight processors.
Astrograd ships with validated modules for the orbital environment, rigid body dynamics, ephemerides, frame and time conversions, numerical integration and finite-state machines needed to build sophisticated spacecraft simulators for control system development and verification.
Astrograd is released under a permissive open-source license for the aerospace community, together with representative examples and tutorials covering hardware accelerated robust satellite control design, frequency-domain analysis, re-entry corridor analysis, uncertainty propagation, trajectory optimization and embedded code generation.
Speaker:
Dr. Lorenzo Capra is a Research Fellow at the European Space Agency, where he develops advanced Guidance, Navigation and Control methods for in-orbit servicing and proximity operations, combining trajectory optimization and machine learning with hardware-in-the-loop validation in ESA’s GRALS robotic test facility. He previously worked as a GNC Engineer at Astroscale Japan, leading the relative navigation architecture for future IOS missions. He earned his PhD from Politecnico di Milano, focusing on AI-based GNC simulation frameworks for orbital assets.
The seminar will be streamed at this link https://politecnicomilano.webex.com/meet/stefano.silvestrini
