We are building a custom Star Tracker. The absolute core bottleneck of our system is processing camera images and matching the detected stars to a massive 3D star catalog database in real-time.
About the Role: We need a self-driven intern to research, design, and implement highly optimized Predictive Catalog Lookups and Attitude Estimation algorithms. This role focuses on minimizing computational complexity: using the prior frame's attitude to tightly bound the database search space, matching patterns rapidly, and updating the satellite's 3D orientation matrix/quaternion
.
Responsibilitie
- s:Design and implement highly optimized spatial database indexing structures (e.g., k-d trees, k-vector lookups, or localized spherical grids) to drastically reduce star search time
- s.Develop a predictive tracking pipeline that uses the satellite's previous attitude estimate to calculate a tiny Window of Interest (WOI) in the star catalog, bypassing global searche
- s.Build robust star pattern-matching logic that handles spatial noise, missing stars, and false centroids from sensor
- s.Estimate orientation to compute the updated attitude from the matched vector
s.
Qualificatio
- ns:Strong grasp of data structures and algorithmic complexity (O(N) vs O(log N) or O(1) optimizatio
- n).Solid foundation in linear algebra, 3D rotations (Quaternions, Matrix transformations), and coordinate geomet
- ry.Proficiency in Python (for rapid prototyping) or C++ (for embedded executio
- n).A self-starter who can dig into research papers on fast star identification techniques and code them from scrat
ch.
