Interactive supplement and reproducibility hub for the bioRxiv preprint Speed-Dependent Turning Strategies in Quadrupedal Locomotion: Insights from Computational Modeling. The paper extends a previously published quadrupedal locomotion model to compare three turning asymmetries, measure turning curvature and inner/outer duty-factor ratios across walking speeds, and test how axial foot-placement shifts expand the stable turning range.
Research group: Yaroslav Molkov's group at Georgia State University
Figure 7 compares the maximum stable curvature produced by body bending, lateral force, and lateral limb shift across walking speeds under the strategy-specific axial-shift settings used in the paper.
With axial shift used as stride control, body bending is the strongest low-velocity strategy, especially below about 5 cm/s, where axial flexion can reorient the body while remaining statically stable.
At roughly 5 to 15 cm/s, forelimb-applied lateral force produces the largest maximum stable curvature in the model, outperforming body bending once turning demands become more dynamic.
At higher speeds, the best-performing strategy is a mediolateral shift of limb touchdown positions. In the model, that foot-placement asymmetry preserves maneuverability as roll-over risk grows.
The paper tracks inner/outer duty-factor ratios and finds a consistent steering-related asymmetry in the forelimbs across strategies, while hindlimb duty factors reorganize differently for body bending, lateral force, and lateral limb shift.
The paper extends the Molkov et al. (2024) quadrupedal locomotion model to compare three turning asymmetries: body bending (geometric), lateral force applied to the forelimbs (dynamic), and lateral limb shift at touchdown (kinematic). Across walking speeds, the authors evaluate the maximum curvature that can be sustained during stable locomotion and track how turning reshapes inner/outer duty-factor ratios.
A central modeling result is that axial (sagittal) shift of the footfall targets acts as stride control: it enlarges the stable turning envelope for all three strategies, but in strategy-specific ways. Under those optimized conditions, body bending performs best at low speed, lateral force at intermediate speed, and lateral limb shift at higher speed. The paper also tests pairwise combinations, showing that mixed strategies mainly bridge the transition zones between the single-strategy speed regimes.
Explore steering controls in the browser and inspect the turning behaviors without building video files first. This is the main interactive entry point for the companion site.
Launch animatorRead the full paper on bioRxiv or jump directly to the PDF when you want the publication itself rather than the interactive supplement.
Open preprintDownload the project archive containing the C++ model, scripts, and build targets used to regenerate the example media and parameter sweeps.
Download source
Local reproduction requires g++, make, gnuplot, ffmpeg, and bc.
The commands below match the project targets used to build the principal example animations and parameter sweeps.
sudo apt update sudo apt install build-essential gnuplot ffmpeg bc make
make results/lateral_force.mp4 make results/body_bending.mp4 make results/lateral_shift.mp4 make delta make force make shift