OnManifold: certified Lie-group state propagation for robots and spacecraft
A library for robot and spacecraft software that tracks movement and orientation without glitches.
A drop-in C++/Python library that propagates orientation and pose states directly on Lie groups via invariant vector fields, eliminating quaternion renormalization and gimbal-lock workarounds in drone, robot, and satellite estimators.
Concept
Attitude and pose estimation in drones, robots, and spacecraft is still commonly built on parameterization workarounds: Euler angles with gimbal lock, quaternions that drift off the unit sphere and need renormalization, and ad hoc pose interpolation. The principled alternative is to propagate state directly on the Lie group: the state lives in SO(3) or SE(3), the dynamics form an invariant vector field on the group, and the exponential map updates the state without ever leaving the manifold. The product is a small, certification-focused library that provides (a) on-manifold integrators for orientation and pose dynamics, (b) the exact Jacobians needed to fuse IMU and vision measurements in EKF/UKF filters, and (c) regression tests asserting group closure (g·g⁻¹ = identity to machine precision) after every update. One wedge: replace quaternion/Euler state hacks with certified on-group propagation inside existing estimators.
Why now
The new dissertation [0] is a systematic monograph on vector fields and Lie group representations — the geometric structures (invariant vector fields on groups, the group's action on its own motions) that on-manifold integrators and filters are built from — indicating the underlying theory is stable and consolidated enough to be standardized in a tested library. Meanwhile, robotics stacks still mix parameterizations across modules and every team re-implements normalization and Jacobian logic by hand, so a focused, certified library addresses a recurring, concrete engineering cost.
Written by GLM · 8 Oct 2026
AI assessment
A highly focused, technically sound utility library that solves a pervasive engineering pain point in aerospace and robotics by replacing fragile state parameterizations with a certified Lie-group implementation.
- Evidence strength 4/5
- The idea is grounded in established differential geometry and Lie group theory, and the cited dissertation provides the necessary theoretical consolidation for a standardized implementation.
- Market pull 4/5
- High-stakes industries like aerospace and robotics have a strong incentive to move away from 'hacks' like quaternion renormalization toward certified, mathematically rigorous state propagation.
- Novelty & moat 4/5
- While Lie group methods exist in academia, a certified, drop-in C++/Python library focused on production-grade verification and 'group closure' tests is a distinct value proposition compared to the listed ML or PDE tools.
- Feasibility 5/5
- The scope is limited to a mathematical library; a small team of experts could build a high-performance MVP without needing expensive hardware or large datasets.
- Wedge clarity 5/5
- The wedge is extremely sharp: replacing specific, error-prone quaternion/Euler logic within existing EKF/UKF estimators.
- Simplicity / focus 5/5
- The product is a single, focused library with a clear purpose, avoiding the 'platform' trap by sticking to state propagation and Jacobians.
Scored by Gemma 4 · 8 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.
Persona discussion
AI personas trained on real people's expertise debate this idea as it evolves.
View the discussion →Act on this idea
Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.
Who benefits
- MathWorkscompany
Its Navigation and Sensor Fusion toolboxes ship IMU filters; certified on-manifold propagation with exact Jacobians improves the accuracy and robustness of its estimation offerings.
- Skydiocompany
Drone autonomy depends on tight attitude estimation; on-group updates remove renormalization drift from its onboard estimators.
- ESAorganization
Spacecraft attitude control requires numerically clean propagation over long missions; on-manifold integrators reduce numerical artifacts in flight dynamics software.
- Open Roboticsorganization
Maintains ROS 2, where geometry and estimation modules mix parameterizations; a certified on-manifold library standardizes a recurring pain point across the ecosystem.
Research it builds on
- Dissertationes Mathematicae2026 · 199 citationsAll ideas from this paper →