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

Hojun Jang; Jinseok Bae; Donggeun Lim; Taesun Kim; Seungbin Jeong; Younghwan Lee; Yong Seok Kim; Young Min Kim

cs.CVOct 28, 2026

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

Authors: Hojun Jang, Jinseok Bae, Donggeun Lim, Taesun Kim, Seungbin Jeong, Younghwan Lee, Yong Seok Kim, Young Min Kim

Organizations: Dept. of Electrical and Computer Engineering, Seoul National University · Interdisciplinary Program in Artificial Intelligence, Seoul National University · Institute of New Media and Communications, Seoul National University

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? We introduce StepControl, an autoregressive diffusion model conditioned on a future binary left/right foot-contact schedule, without requiring target root and body poses, or semantic prompts. StepControl represents contact as explicit temporal events and injects them through dedicated cross-attention with classifier-free guidance. We further introduce a stochastic mixture noise schedule and warmup strategy for stable long-horizon autoregressive generation. Experiments show that StepControl follows prescribed contact timing more accurately than adapting existing spatially guided motion generators to contact constraints. It maintains effective contact control over long continuous rollouts and responds to edited schedules such as cadence changes, stop-and-go pauses, limping gaits, and bilateral jumping. Our results show that minimal foot-contact schedules provide an effective and intuitive temporal interface for controllable motion generation.

Figures & tables

Figure 1: StepControl generates full-body motion from foot-contact schedules without future foot positions or root trajectories. Beyond reproducing ordinary walking cadence, StepControl follows atypical asymmetric schedules.

Figure 2: Overview of StepControl. (a) Motion tokens with per-frame diffusion levels and frame embeddings attend to event condition tokens via dedicated per-layer cross-attention, predicting clean motion $\hat{\mathbf{X}}^{(0)}$ for continuous long-horizon generation. (b) Within each sliding window, 2D foot-contact signals $\mathbf{C}$ are expanded into 9D event-aware representations $\mathbf{E}$ . Conditioned on $\mathbf{E}$ , the denoiser steps down noisy future latents via per-frame DDIM updates ( $k_{i}\to k_{i}^{\prime}$ ), commits the earliest finalized frame $\hat{\mathbf{x}}_{t}^{(0)}$ to clean history, and appends a new terminal noise state ( $k=99$ ) as the window slides forward.

Rollout Horizon	Method	Bal. Acc. (%) $\uparrow$	Tr. F1 (%) $\uparrow$	Slip (mm/f) $\downarrow$	RA-LB (mm) $\downarrow$	Mean Pen. (mm) $\downarrow$
1-Second	DART	74.34	15.26	7.84	109.97	0.30
	ARDY	69.37	16.92	20.69	158.62	0.47
	StepControl	79.90	36.15	17.63	144.25	0.00
	StepControl (+Opt.)	80.60	33.59	8.64	144.25	0.00
5-Second	DART	69.25	16.52	3.28	180.68	0.31
	ARDY	64.28	10.29	13.52	185.86	0.34

Table 1: Quantitative comparison on autoregressive motion generation across extended horizons. We evaluate contact synchronization and progressive rollout stability against autoregressive baselines (DART [ 38 ] , ARDY [ 39 ] ) across LAFAN1 [ 10 ] test sequences.

Figure 3: 20 s Motion Generation Result of StepControl. StepControl generates natural and consistent long-horizon motion from only foot-contact schedule.

Figure 4: Qualitative comparison on contact schedule editing (bilateral jump). StepControl better synchronizes simultaneous flight and landing phases, whereas DART [ 38 ] fails to coordinate flight and ARDY [ 39 ] exhibits noisy contact timing and unstable poses. ARDY is rendered with its native body model.

Modification	Method	Bal. Acc. $\uparrow$ (%)	Tr. F1 $\uparrow$ (%)	Slip $\downarrow$ (mm/f)
Cadence Scaling	DART	67.91	9.13	3.44
	ARDY	77.41	16.44	6.19
	StepControl	76.11	33.57	16.49
Stop-and-Go	DART	69.49	21.13	2.80
	ARDY	78.88	16.81	4.05
	StepControl	82.18	37.31	11.77

Table 2: Quantitative evaluation of interactive contact scheduling (5 s horizon). We evaluate responsiveness to temporal cadence scaling, gait transitions, and atypical gait synthesis against DART [ 38 ] and ARDY [ 39 ] across LAFAN1 [ 10 ] test sequences.

Ablation Configuration	Bal. Acc. (%) $\uparrow$	Tr. F1 (%) $\uparrow$	Slip (mm/f) $\downarrow$	RA-LB (mm) $\downarrow$	Bo. Jerk (mm/f 3 ) $\downarrow$
(a) Contact Feature Representation
Raw 2D Contact	76.23	21.79	16.06	182.54	77.92
Delta 5D Contact	79.30	23.79	17.45	175.75	75.48
Event 9D Contact (Ours)	79.30	39.58	15.54	181.64	70.54
(b) Training Dataset
LAFAN1 Only	80.99	26.00	21.02	180.43	83.34

Table 3: Ablation studies on the LAFAN1 [ 10 ] test benchmark. We evaluate 5-second prefixes of the same 20-second evaluation segments used in the main benchmark.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Method	Per-Frame	20 s Rollout	Throughput
DART [ 38 ] (50-step Opt.)	$\approx 2267\,\text{ms}$	$\approx 907\,\text{s}$	$0.44\,\text{fps}$
ARDY [ 39 ] (10-step Diff.)	$\approx 75.0\,\text{ms}$	$\approx 30.0\,\text{s}$	$13.3\,\text{fps}$
StepControl (Ours)	$\mathbf{7.5\,\text{ms}}$	$\mathbf{2.9\,\text{s}}$	$\mathbf{133.3\,\text{fps}}$

Appendix

Table A1 : Inference runtime and generation throughput comparison for 20-second (400-frame) rollouts on a single NVIDIA RTX 4090 GPU. StepControl achieves real-time throughput (

>130

fps) via direct single-pass generation.

Loss Formulation	Bal. Acc. (%) $\uparrow$	Tr. F1 (%) $\uparrow$	Slip (mm/f) $\downarrow$	RA-LB (mm) $\downarrow$	Bo. Jerk (mm/f 3 ) $\downarrow$
Only Motion Reconstruction Loss ( $\mathcal{L}_{\text{mse}}$ )	73.87	21.60	22.40	204.35	71.45
w/o Stance Velocity Loss ( $\mathcal{L}_{\text{stance}}$ )	76.84	32.27	19.93	201.03	70.94
w/o Foot Height Loss ( $\mathcal{L}_{\text{height}}$ )	78.02	33.30	15.39	180.55	70.68
w/o 4D Contact MSE Loss ( $\mathcal{L}_{\text{contact4}}$ )	79.03	34.21	15.80	181.17	70.28
Full Multi-Task Loss (StepControl)	79.30	39.58	15.54	181.64	70.54

Appendix

Table A2 : Ablation on loss formulations (LAFAN1 test benchmark). All variants use the same mixture noise schedule, pretraining initialization, fine-tuning protocol, evaluation segments, and random seed.

CFG Guidance Scale	Bal. Acc. (%) $\uparrow$	Tr. F1 (%) $\uparrow$	Slip (mm/f) $\downarrow$	RA-LB (mm) $\downarrow$	Bo. Jerk (mm/f 3 ) $\downarrow$
$s=0.0$ (Unconditional Baseline)	52.08	3.90	38.36	208.98	69.12
$s=1.0$ (Standard Conditional Prior)	74.67	24.52	16.91	183.16	69.49
$s=2.0$ (Default StepControl)	79.30	39.58	15.54	181.64	70.54

Appendix

Table A3 : Ablation on classifier-free guidance (CFG) scale

s

(LAFAN1 test benchmark). Increasing

s

enhances temporal contact adherence and transition precision.

Inference	Bal. Acc. (%) $\uparrow$	Tr. F1 (%) $\uparrow$	Slip (mm/f) $\downarrow$	RA-LB (mm) $\downarrow$	Bo. Jerk (mm/f 3 ) $\downarrow$
Chunk AR	76.32	30.75	19.96	203.63	75.59
Progressive (Ours)	79.30	39.58	15.54	181.64	70.54

Appendix

Table A4 : Rollout strategy comparison on the LAFAN1 test benchmark.

	Low	Mid	High
Transition rate (events/s)	1.2–3.2	3.4–4.0	4.4–8.0
Windows	20	16	16
Target events	253	300	460
DART	17.82	15.93	16.88
ARDY	13.42	16.64	18.54
Raw 2D Contact	26.55	33.34	25.65

Appendix

Table A5 : Transition F1 (%) by contact-transition rate.

Human Motion Generation

Human motion generation creates articulated movement from language, audio, video, spatial constraints, or interaction goals.

See all 292 papers

Training-free Controllable Human Motion Generation under Heterogeneous Constraints
Jul 2, 2026Xiaofei Hui, Bo Yan, Haoxuan Qu +2Constrained Generative Modeling Human Motion Generation
Lancaster University, United Kingdom · Shenzhen International Graduate School, Tsinghua University, China
Diffusion Path Alignment for Long-Range Motion Generation and Domain Transitions
Mar 31, 2026Haichao Wang, Alexander Okupnik, Yuxing Han +3Human Motion Generation Diffusion Models for Motion Generation
Shenzhen International Graduate School, Tsinghua University, China · Institute of Information Systems, Universität Liechtenstein, Liechtenstein · Department of Computer Science, New York University, USA +1
Kimodo: Scaling Controllable Human Motion Generation
Mar 16, 2026Davis Rempe, Mathis Petrovich, Ye Yuan +21Human Motion Generation Diffusion Models for Motion Generation
NVIDIA

Jul 2, 2026J/K move · Enter open · S save

Training-free Controllable Human Motion Generation under Heterogeneous Constraints

Xiaofei Hui, Bo Yan, Haoxuan Qu, Hossein Rahmani+1

Lancaster University, United Kingdom · Shenzhen International Graduate School, Tsinghua University, China

Constrained Generative Modeling Human Motion Generation Diffusion Models for Motion Generation

Open paper AskPDF arXiv

Training-free controllable motion generation has attracted growing interest for enabling flexible constraint enforcement without constraint-specific training. However, existing training-free methods require constraints to be continuous objective-based with differentiable losses, while many real-world requirements are criterion-based and provide only discontinuous, sparse, or even black-box feedback. In this paper, we propose Motion-Inference-as-Control (MIC), the first training-free motion generation framework that handles both continuous objective-based and criterion-based motion constraints under a shared mechanism. The key idea is to cast diffusion-based motion generation as a stochastic control problem. This perspective not only provides principled and practically effective step-wise control laws that support criterion-based constraints without requiring differentiability and naturally accommodate objective-based constraints as a special case, but also motivates a control-oriented constraint coordination mechanism that adaptively balances and reconciles motion constraints during generation. Experiments across diverse constraint settings demonstrate the effectiveness of our framework.

Abstract

Human Motion Generation

Explore similar work

Training-free Controllable Human Motion Generation under Heterogeneous Constraints

Diffusion Path Alignment for Long-Range Motion Generation and Domain Transitions

Kimodo: Scaling Controllable Human Motion Generation