Problem
Research questions and friction points this paper is trying to address.
legged robots
proprioceptive odometry
contact assumption
state estimation
observability
Interacting Multiple Model Proprioceptive Odometry for Legged Robots
🤖 AI Summary
This work addresses the degradation in accuracy and state drift of proprioceptive odometry for legged robots when external sensing fails and the point-contact assumption is violated. To overcome these limitations, the authors propose a proprioceptive odometry framework based on the Interacting Multiple Model (IMM) approach. By moving beyond the restrictive point-contact assumption, the method incorporates multiple contact models and enables online mode switching and fusion within a unified probabilistic framework. Integrating legged robot kinematics with probabilistic state estimation, the proposed approach significantly enhances the robustness and accuracy of pose estimation under complex contact conditions. Both simulation and real-world experiments demonstrate that the method achieves notably higher pose estimation accuracy than existing solutions while maintaining computational efficiency.
📝 Abstract
State estimation for legged robots remains challenging because legged odometry generally suffers from limited observability and therefore depends critically on measurement constraints to suppress drift. When exteroceptive sensors are unreliable or degraded, such constraints are mainly derived from proprioceptive measurements, particularly contact-related leg kinematics information. However, most existing proprioceptive odometry methods rely on an idealized point-contact assumption, which is often violated during real locomotion. Consequently, the effectiveness of proprioceptive constraints may be significantly reduced, resulting in degraded estimation accuracy. To address these limitations, we propose an interacting multiple model (IMM)-based proprioceptive odometry framework for legged robots. By incorporating multiple contact hypotheses within a unified probabilistic framework, the proposed method enables online mode switching and probabilistic fusion under varying contact conditions. Extensive simulations and real-world experiments demonstrate that the proposed method achieves superior pose estimation accuracy over state-of-the-art methods while maintaining comparable computational efficiency.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Interacting Multiple Model
Proprioceptive Odometry
Contact Hypothesis
Legged Robots
State Estimation
🔎 Similar Papers
No similar papers found.
