Slip decomposition and state estimation for rovers
Year: 2026
Type: article
Abstract: • Proposes a novel decomposition of wheel slip: conflicting and compatible types • Presents three estimation frameworks addressing the two slip components • Kinematic estimator fuses wheel data to resolve conflicting slip • Dynamic estimator uses terramechanics to handle compatible slip • Combined method improves state estimation, validated by experiments Odometry is a widely used state-estimation technique for exploration rovers, determining rover position and velocity from observed wheel motions. However, wheel slip remains a primary source of error, particularly on soft terrain, leading to significant inaccuracies during travel. Previous studies on wheel-slip modelling have predominantly focused on the interaction between a single wheel and the terrain, often failing to capture the complexities of multi-wheel systems in real-world scenarios. This paper presents a wheel-encoder framework for rover state estimation that integrates rover kinematics with multi-wheel coupling and wheel-soil interaction dynamics. The first objective is to introduce a novel definition of wheel-slip decomposition, dividing slip into two components: one addressing internal kinematic conflicts among wheels and another arising from different wheels collectively satisfying the rover’s kinematic constraints. The second objective is to develop three estimation frameworks operating at both kinematic and dynamic levels to accurately resolve the defined slip components. The proposed frameworks are validated through experiments using a rover prototype. Experimental results demonstrate that the proposed method reduces distance and orientation errors from 4–6% to below 1%, achieving more than an 80% improvement in accuracy compared with classical odometry.
Source: Mechanism and Machine Theory
Authors: Qingfeng Lou, József Kövecses
Institutions: McGill University, Obuda University
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,060
OpenAlex ID: W7128474747