Interacting Multiple Model Proprioceptive Odometry for Legged Robots

TL;DR

This paper introduces an IMM-based proprioceptive odometry method for legged robots that improves pose estimation accuracy by handling multiple contact scenarios and mode switching, especially when exteroceptive sensors are unreliable.

Contribution

It presents a novel probabilistic framework that fuses multiple contact hypotheses for better state estimation during legged robot locomotion.

Findings

Achieves superior pose accuracy compared to existing methods.

Maintains real-time computational efficiency.

Effective in real-world and simulated environments.

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.