Inertial Steering and Motion Prediction of Tracked Vehicles on Low-Adherence Surfaces

Mathis Turgeon-Roy; Mohamed Ounally; Veronica Vannini; Olivier Roy; François Pomerleau

cs.ROOct 28, 2026

Inertial Steering and Motion Prediction of Tracked Vehicles on Low-Adherence Surfaces

Authors: Mathis Turgeon-Roy, Mohamed Ounally, Veronica Vannini, Olivier Roy, François Pomerleau

Organizations: Northern Robotics Laboratory (Norlab), Université Laval, Québec, QC, Canada · TSCF, INRAE, Clermont-Ferrand, France

Abstract

Tracked vehicles are widely used in off-road applications, but their conventional motion models often rely on no-slip assumptions that become inaccurate on low-adherence surfaces. Rather than treating inertial effects arising from slip as a disturbance to be minimized, this work investigates how inertial effects arising from a moving mass can be deliberately exploited for steering on low-friction surfaces. A dynamic model for an articulated tracked vehicle with a passive trailer is developed, allowing inertial effects from trailer motion to be captured. The approach is experimentally evaluated on ice, where our dynamic model outperforms the kinematic baseline. We also propose a mixed dynamic-kinematic model to adapt predictions to both surface and situation by exploiting predictions from both methods. Finally, we demonstrate inertial steering on a mono-track vehicle without a conventional steering mechanism, showing that controlled movements of an internal mass can generate turning motion on ice. The resulting behavior can be predicted using the proposed dynamic model by representing the internal mass as a frictionless trailer, revealing the analogous dynamics between the two configurations. These results demonstrate the feasibility of inertial steering as an alternative steering mechanism for tracked vehicles operating in low-adherence environments.

Figures & tables

Fig. 1: Motion of the MTT-154 on high- and low-adherence ground. (a) On grass, the articulated vehicle follows the classic kinematic model ICR computation, which depends only on the track speed and the articulation angle. (b) On ice, the angular acceleration of the trailer mass becomes relevant to the motion of the articulated vehicle. (c) The same inertial effect steers a mono-track, which has no steering mechanism, through an actuated internal mass. Low friction degrades kinematic models but opens the opportunity for a steering mechanism based on inertia.

Fig. 2: Planar model of the two configurations: (a) Articulated vehicle: the actuated hitch, at a distance

d_{x}

behind

Q

, swings a passive wheeled trailer of mass

m

, whose wheels receive the lateral ground force

P

, perpendicular to the trailer. (b) Mono-track: an actuated arm pivoting also at a distance

d_{x}

behind the main-body center of mass

Q

swings the mass

m

, which shifts the COM of the vehicle. In both cases, a mass

m

lies at a distance

L

from the pivot, and the articulation angle

\varphi

is measured from

-x

, positive counterclockwise. The body twist

\bm{\xi}=(v,w,\omega)

is expressed in the body frame, with axes

x

and

y

, attached to

Q

. Both configurations follow the same equations of motion, except for the trailer ground force, which is zero for the mono-track, and the trailer weight, which does not load the track.

Fig. 3: Illustration of the discrete track contact and load-distribution model from three orthogonal views. (a) Bottom view (

x

y

) showing the contact points along the track and the corresponding friction forces

\bm{F}_{i}

, computed from

F_{z,i}

, the contact-point position

\bm{r}_{i}

relative to

Q

, and the track speed

v_{t}

. (b) Front view (

y

z

) showing the lateral eccentricity

e_{y}

and the resulting distribution of the total normal load

F_{g}

between the two contact strips, with strip reactions

F_{s,1}

and

F_{s,2}

of Eq. 5 . (c) Side view (

x

z

) showing the longitudinal eccentricity

e_{x}

and the resulting continuous line-load distribution

q(x)

, from which the discrete contact-point loads

F_{z,i}

of Eq. 7 are obtained.

Fig. 4: Simulated trajectory of $10\text{\,}\mathrm{s}$ on ice with the mixed kinematic-dynamic model predictions for different $\lambda$ values. $\lambda=0$ is equivalent to a kinematic model, while $\lambda=1$ corresponds to the predictions of our dynamic model. The trajectory represents a typical aggressive maneuver on ice where kinematic and dynamic models diverge.

Fig. 5: Experimental platform in mono-track configuration with mechanical arm installed and loaded, as well as sensors used to obtain our reference trajectories.

	Characteristic	Symbol	Unit	Mono.	Art.
Physical	Swung mass	$m$	kg	45	156
	Pivot lever length	$L$	m	0.40	1.515
	Pivot–Q longitudinal offset	$d_{x}$	m	0.308	1.624
	Contact patch length	$\ell$	m	1.20
	Contact points per strip	$N_{s}$	–	9
	Contact points	$N$	–	18

TABLE I: Physical parameters of the platform in the mono-track (Mono.) and articulated (Art.) configurations.

	Ice		Grass
Parameter	Track	Trailer	Track	Trailer
$B$	$6.20$	$2.10$	$1.32$	$2.80$
$C$	$1.95$	$1.92$	$1.97$	$1.61$
$E$	$1.91$	$3.27$	$-1.13$	$-0.30$
$\mu$	0.1		0.6

TABLE II: Pacejka tire model parameters of the track model and trailer fitted on ice and grass. The peak factor

D

is not fitted, it follows Eq. 3 .

Fig. 6: RPE in position (left column) and yaw heading (right column) of the dynamic and kinematic models over

3.4\text{\,}\mathrm{km}

of trajectories performed on grass and

1.5\text{\,}\mathrm{km}

on ice. The first row shows the results for driving on grass, while the second one shows results on ice. The full lines represent the median error, while the shaded areas represent the IQR . Note that the scale of RPE for both terrains is an order of magnitude different.

Fig. 7: Share of

5\text{\,}\mathrm{s}

windows in which each

\lambda

gives the lowest position prediction error, over the full grass and ice articulated datasets, where

\lambda=0

is the kinematic model alone and

\lambda=1

the dynamic model alone. Fig. 4 shows the trajectories these two ends predict.

Fig. 8: RPE of the dynamic model against the constant-velocity baseline over recordings of

30

trajectories totaling

818\text{\,}\mathrm{m}

with mono-track configuration on ice performed with the experimental protocol described in Section V-C . The full lines represent the median error, while the shaded areas represent the IQR .

Fig. 9: Mono-track trajectories on the lidar map of the ice rink, with the boards and surrounding structures coloured by height. The reference (dashed) is compared with the predictions of the dynamic model and of the constant velocity baseline over the whole run. (a) The vehicle finishes its curve, straightens its trajectory, and continues in steady forward motion. (b) The vehicle keeps turning with high side slip and finishes in steady circular motion. The constant velocity prediction is shown over its first $10\text{\,}\mathrm{s}$ , after which it keeps circling.

Vehicle Dynamics

Vehicle dynamics studies how vehicles move and respond to forces from tires, terrain, aerodynamics, and control inputs.

See all 79 papers

Sep 18, 2026

When to Waddle: A Comparative Study of Bipedal Torso-Stabilization on Low-Friction Surfaces

Low-friction surfaces challenge bipedal locomotion by limiting the contact forces available during stepping. Inspired by penguin waddling, we investigate how lateral torso motion and center of mass (COM) placement affect locomotion as surface friction changes. Using a five-actuator biped, we compare an upright-gait strategy with a penguin-inspired torso-over-stance-leg strategy across multiple COM placements in simulation and hardware. In the 3-D simulator MuJoCo, we sweep through sinusoidal leg and hip actuation parameters across four friction coefficients mu = 0.1, 0.3, 0.5, 0.7. In simulation, torso-over-stance-leg motion produces more successful controllers and higher forward speeds at low friction, with the highest speed occurring for the high-COM configuration. Hardware experiments show the same low-friction speed trend: at mu=0.12, torso-over-stance-leg motion increases forward speed and reduces cost of transport at both tested COM ratios, and the higher COM also improves both measures. The high-COM penguin configuration is the fastest and most energy efficient while maintaining low sideways foot motion. At mu=0.45, the COM trend reverses: the lower-COM configurations are faster and more energy efficient, while gait strategy has little effect on forward speed but still changes sideways foot motion. These results show that the effects of lateral torso motion and COM placement depend on the available friction, and that forward speed, energy use, and slip-related foot motion can be modulated with a penguin-inspired torso motion on hardware.

Ben Gu, Naomi Oke, George Ortiz +4

Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA · Department of Mechanical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA · Mechanical and Aerospace Engineering, New York University, New York, NY 10012, USA

Humanoid Robot Locomotion Energy-Efficient Robot Locomotion Robot Locomotion

Ask PDF

Sep 20, 2026

Steering Through Contact: A Finite-Support Motion Model for Single-Track Center-Articulated Robots

Trajectory planning and control in field robotics rely on predicting how propulsion and steering affect vehicle motion when contact points undergo slip. For articulated vehicles, the point-contact kinematic model (PCK) accounts for the linkage geometry but neglects the rotational resistance distributed along the contacts. We propose a finite-support quadratic model (FSQ) for single-track, center-articulated vehicles that incorporates this resistance through a quasi-static balance of lateral slip. Our approach generalizes the standard PCK formulation by relaxing the contact-point assumption. An exact reduction of the quadratic slip cost to contact moments gives a compact closed-form solution for real-time prediction of lateral velocity and yaw rate. We evaluate the proposed method in real-world experiments across asphalt, grass, ice, and mixed routes, using more than 7 km of data. For five-second predictions, FSQ reduces the weighted median translation and yaw errors by 53.5% and 68.9%, respectively, relative to PCK.

Mohamed Dhia Ounally, Nicolas Samson, Mathis Turgeon-Roy +3

Norlab, Universit´e Laval, Qu´ebec, QC, Canada · TSCF, INRAE, Clermont-Ferrand, France

Robot Motion Prediction Robotics Vehicle Dynamics Trajectory Prediction

Ask PDF

Sep 30, 2026

Occlusion-Aware, Quasi-Static, Stability-Oriented Trajectory Planning on Uneven Terrain

Autonomous navigation in unstructured off-road environments requires reasoning about both vehicle--terrain interaction and environmental unknowns. We propose a model-based framework for generating quasi-static, stability-oriented reference trajectories for rigid, non-articulated four-wheeled vehicles on highly uneven terrain. Our work makes three primary contributions. First, we model blind spots caused by terrain occlusion as coverage-induced epistemic uncertainty in a fixed-feature Fourier terrain representation, quantified through a regularized inverse-Hessian estimate. Second, we propagate this uncertainty through the Nonlinear Least-Squares (NLS) pose/contact model using implicit differentiation and incorporate the resulting pose, contact-point, and per-wheel surface-normal uncertainty terms into trajectory optimization based on the Cross-Entropy Method (CEM). Third, we introduce a Flow Matching model that warm-starts terrain fitting, and we evaluate its fitting-accuracy--latency trade-off while retaining model-based refinement. Across six synthetic terrains with 30 matched start--goal pairs per terrain, the complete framework produced an observed failure rate of 18.9%, compared with 46.1% and 41.7% for two representative baselines and 34.4% for an ablation that removed the propagated-uncertainty scoring. Hardware evaluations span six distinct outdoor environments, with two representative executions presented in the paper and four additional executions included in the supplementary video. The evaluation also reports the accuracy--latency trade-off for the Flow Matching warm start.

Amith Manoharan, Chinmay Mundane, Aayush Bahukhandi +3

Institute of Technology, University of Tartu, Estonia · RRC, IIIT Hyderabad, India · Czech Technical University, Prague

Trajectory Optimization Off-Road Robot Navigation Trajectory Generation Terrain-Aware Robot Locomotion Motion Planning

Ask PDF

Abstract

Vehicle Dynamics

Explore similar work

When to Waddle: A Comparative Study of Bipedal Torso-Stabilization on Low-Friction Surfaces

Steering Through Contact: A Finite-Support Motion Model for Single-Track Center-Articulated Robots

Occlusion-Aware, Quasi-Static, Stability-Oriented Trajectory Planning on Uneven Terrain