🔬 Projects & Code

A curated list of open-source projects, frameworks, and implementations developed by the HEAT community.

Spacecraft Pose Estimation Framework

Author: Julien Posso (Postdoctoral Researcher @ ONERA | PhD Polytechnique Montréal)

Tags: Deep Learning Space Robotics Embedded AI FPGA Model Quantization Real-Time Inference

An end-to-end framework for the design, quantization-aware training (QAT), and multi-platform (CPU, GPU, FPGA) deployment of Neural Networks dedicated to non-cooperative spacecraft pose estimation. It bridges the gap between high-level AI models and real-time orbital edge computing.

📄 Related Publications:

• Mobile-URSONet: an Embeddable Neural Network for Onboard Spacecraft Pose Estimation (ISCAS 2022)

• Real-Time Spacecraft Pose Estimation Using Mixed-Precision Quantized Neural Network on COTS Reconfigurable MPSoC (NEWCAS 2024)

• Real-Time Spacecraft Pose Estimation Using Deep Learning: From Single Images to Video Sequences (Under review at IEEE TAES)

🔗 Explore the framework on GitHub

---

[Add your project here]

Have a repository showcasing an elegant Chisel IP, a custom TensorRT deployment, or a novel hardware architecture? Click the ✏️ edit icon and add it following the format above.