Reproduce “Legged Robot State-Estimation Through Combined Forward Kinematic and Preintegrated Contact Factors” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ross Hartley, Josh Mangelson, Lu Gan, Maani Ghaffari Jadidi, Jeffrey M. Walls, Ryan M. Eustice, Jessy W. Grizzle
A state-estimation method for legged robots integrates forward kinematic and preintegrated contact factors into a factor graph framework, reducing reliance on visual sensors. This approach significantly reduces positional drift and improves robustness, achieving an end-to-end translation error of approximately 1.3 meters over a 100-second real-world robot trajectory.
Key insights
- Leg odometry, when properly modeled and integrated into a factor graph, provides a robust and complementary source of information to IMUs, reducing drift and enhancing state estimation, especially in the absence of reliable visual data.
- Preintegration techniques, typically applied to IMUs, can be effectively adapted for contact measurements, allowing for efficient fusion of high-frequency leg odometry data within the factor graph.
- By fusing diverse sensor modalities (IMU, joint encoders, contact sensors) probabilistically, the framework can more accurately estimate IMU biases and maintain robust state estimation without relying on external visual inputs for absolute corrections.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.