🤖 AI Summary
Reconstructing kinematically consistent vehicle state sequences (position, velocity, acceleration, heading) from noisy, position-only trajectory observations—while enforcing integral relationships (e.g., velocity as the integral of acceleration) and directional motion constraints (i.e., motion strictly along the heading direction).
Method: We propose an end-to-end trajectory reconstruction framework that jointly models B-spline parameterization, physics-based regularization (embedding numerical integration and orientation-driven motion constraints), and supervised feature regression.
Contribution/Results: To our knowledge, this is the first approach unifying these three components in a single differentiable architecture. We release a configurable Python library implementing the method. Experiments on real-world data demonstrate significant improvements in trajectory accuracy—yielding reconstructions substantially closer to ground truth—and enhanced performance in downstream machine learning tasks reliant on high-fidelity reference trajectories.
📝 Abstract
This article presents a method for estimating the dynamic driving states (position, velocity, acceleration and heading) from noisy measurement data. The proposed approach is effective with both complete and partial observations, producing refined trajectory signals with kinematic consistency, ensuring that velocity is the integral of acceleration and position is the integral of velocity. Additionally, the method accounts for the constraint that vehicles can only move in the direction of their orientation. The method is implemented as a configurable python library that also enables trajectory estimation solely based on position data. Regularization is applied to prevent extreme state variations. A key application is enhancing recorded trajectory data for use as reference inputs in machine learning models. At the end, the article presents the results of the method along with a comparison to ground truth data.