InEKFormer: A Hybrid State Estimator for Humanoid Robots

2025-11-20

AI Summary

Accurate, low-latency real-time estimation of the floating-base state is critical for stable dynamic bipedal locomotion of humanoid robots in complex environments. Conventional filtering approaches—such as Invariant Extended Kalman Filters (InEKF)—rely on manual parameter tuning and suffer from limited robustness, while purely data-driven methods lack generalizability. To address these limitations, we propose an InEKF-Transformer hybrid state estimation framework that tightly integrates the geometric-aware modeling capability of InEKF with the Transformer’s long-range temporal modeling and autoregressive learning capacity, enabling end-to-end trainable, geometry-consistent state estimation. Evaluated on the real-world RH5 humanoid robot dataset, our method significantly outperforms both standard InEKF and KalmanNet in pose, position, and velocity estimation accuracy, and demonstrates superior robustness under motion disturbances. This advancement provides a more reliable perception foundation for dynamic bipedal control.

Technology Category

• Intelligent Robots: State Estimation
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Humanoid robots have great potential for a wide range of applications, including industrial and domestic use, healthcare, and search and rescue missions. However, bipedal locomotion in different environments is still a challenge when it comes to performing stable and dynamic movements. This is where state estimation plays a crucial role, providing fast and accurate feedback of the robot's floating base state to the motion controller. Although classical state estimation methods such as Kalman filters are widely used in robotics, they require expert knowledge to fine-tune the noise parameters. Due to recent advances in the field of machine learning, deep learning methods are increasingly used for state estimation tasks. In this work, we propose the InEKFormer, a novel hybrid state estimation method that incorporates an invariant extended Kalman filter (InEKF) and a Transformer network. We compare our method with the InEKF and the KalmanNet approaches on datasets obtained from the humanoid robot RH5. The results indicate the potential of Transformers in humanoid state estimation, but also highlight the need for robust autoregressive training in these high-dimensional problems.

Problem

• Developing a hybrid state estimator for stable humanoid robot locomotion
• Addressing challenges in bipedal movement across different environments
• Improving state estimation accuracy without expert noise parameter tuning

Innovation

• Hybrid method combining InEKF and Transformer network
• Transformer architecture applied to humanoid state estimation
• Autoregressive training for high-dimensional robot problems