StepControl: Long-Horizon Human Motion Generation from Foot-Contact Schedules

WACV 2027

Hojun JangJinseok BaeDonggeun LimTaesun KimSeungbin JeongYounghwan LeeYong Seok KimYoung Min Kim

Seoul National University

Specify when the feet touch the ground,
and generate full-body motion, without a root trajectory or semantic prompt.

Abstract

Controllable human motion generation has largely focused on specifying what action should occur or where the body should move. We study a complementary question: can full-body motion be controlled by specifying only when each foot contacts the ground? StepControl is an autoregressive diffusion model conditioned on a future binary left/right foot-contact schedule, without requiring target root and body poses or semantic prompts. It represents contact as explicit temporal events and injects them through dedicated cross-attention with classifier-free guidance. A stochastic mixture noise schedule and rollout warmup support stable long-horizon generation. StepControl follows prescribed contact timing over long rollouts and responds to edited schedules including cadence changes, stop-and-go pauses, limping gaits, and bilateral jumping.

Method

Overview of the StepControl training and autoregressive generation pipeline

StepControl is an autoregressive diffusion model that generates motion from a short observed prefix and a future foot-contact schedule. (1) Event-aware contact: binary left/right contacts are expanded with onset, release, temporal change, and validity signals. (2) Dedicated conditioning: motion tokens attend to contact tokens at every Transformer layer, and classifier-free guidance controls adherence at inference. (3) Progressive rollout: a mixture noise schedule trains the denoiser for sliding-window generation, and a short warmup initializes the first rollout window.

Results (Baseline Comparisons)

We compare raw StepControl (without post-optimization) against spatial-trajectory baselines DART and ARDY, adapted to foot-contact constraints under identical timing conditioning. ARDY is visualized with its native mesh configuration.

Dance Locomotion

Expressive, rhythmic contact patterns generated without text or audio guidance.

StepControl (Ours)
DART
ARDY

Schedule Editing: Slow Cadence Walk

Low-frequency stepping schedule with lengthened single-stance duration.

StepControl (Ours)
DART
ARDY

Schedule Editing: Stop and Go

Locomotion interleaved with standstill periods, requiring precise stance locking and resumption.

StepControl (Ours)
DART
ARDY

Schedule Editing: Bilateral Jump

Simultaneous release, airborne phase, and coordinated two-foot landing.

StepControl (Ours)
DART
ARDY

Schedule Editing: Limping Gait

Asymmetric left/right stance durations requiring non-uniform gait timing.

StepControl (Ours)
DART
ARDY

Results (Long Horizon, 20 s)

Continuous 20-second rollouts from short observed prefixes, comparing raw StepControl with StepControl + Opt., a lightweight stance-foot velocity correction. All StepControl motions in the baseline comparisons above are raw generations.

Locomotion Sequence 1

StepControl (Raw)
StepControl + Opt.

Locomotion Sequence 2

StepControl (Raw)
StepControl + Opt.

BibTeX

@inproceedings{stepcontrol2027,
  author    = {Jang, Hojun and Bae, Jinseok and Lim, Donggeun and Kim, Taesun and Jeong, Seungbin and Lee, Younghwan and Kim, Yong Seok and Kim, Young Min},
  title     = {StepControl: Long-Horizon Human Motion Generation from Foot-Contact Schedules},
  booktitle = {Proceedings of the Winter Conference on Applications of Computer Vision (WACV)},
  month     = {January},
  year      = {2027}
}