Characterizing gaussian mixture of motion modes for skid-steer state estimation

2025-04-30

AI Summary

To address the significant tire-terrain slippage of slip-steered wheeled mobile robots (SSWMRs) in complex terrain—which degrades motion model fidelity and impedes high-precision state estimation—this paper proposes a dynamic modeling and estimation framework integrating Gaussian Mixture Models (GMM) with the Interacting Multiple Model (IMM) algorithm. We introduce GMM for the first time to perform unsupervised clustering and online identification of SSWMR motion modes, enabling construction of a multi-model set tailored to distinct slippage characteristics; these models are embedded within the IMM framework to achieve real-time, adaptive angular velocity estimation. The method balances modeling accuracy and computational efficiency. Experimental validation on a mid-scale platform demonstrates a 32% reduction in angular velocity estimation error and model-switching latency under 15 ms—substantially outperforming single-model approaches.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Skid-steered wheel mobile robots (SSWMRs) are characterized by the unique domination of the tire-terrain skidding for the robot to move. The lack of reliable friction models cascade into unreliable motion models, especially the reduced ordered variants used for state estimation and robot control. Ensemble modeling is an emerging research direction where the overall motion model is broken down into a family of local models to distribute the performance and resource requirement and provide a fast real-time prediction. To this end, a gaussian mixture model based modeling identification of model clusters is adopted and implemented within an interactive multiple model (IMM) based state estimation. The framework is adopted and implemented for angular velocity as the estimated state for a mid scaled skid-steered wheel mobile robot platform.

Problem

• Modeling skid-steer robot motion with unreliable friction models
• Ensemble modeling for distributed performance and real-time prediction
• Gaussian mixture model for state estimation in skid-steer robots

Innovation

• Gaussian mixture model for motion mode identification
• Interactive multiple model for state estimation
• Ensemble modeling to distribute performance requirements