---
title: "TPPMG: Temporal Planning-driven Progressive Motion Generation"
authors:
  - "Ruoyu Wang"
  - "Can Deng"
  - "Xinyi Li"
  - "Zhuo Li"
doi: "10.24963/ijcai.2026/197"
url: "https://ijcai.org/proceedings/2026/197"
pdf: "https://ijcai.org/proceedings/2026/0197.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "1767-1775"
track: "Main Track"
keywords:
  - "Computer Vision: Motion and tracking"
  - "Computer Vision: Action and behavior recognition"
  - "Computer Vision: 3D computer vision"
---

# TPPMG: Temporal Planning-driven Progressive Motion Generation

Ruoyu Wang, Can Deng, Xinyi Li, Zhuo Li

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 1767-1775. DOI: [10.24963/ijcai.2026/197](https://doi.org/10.24963/ijcai.2026/197)

## Abstract

Most text-to-motion models treat motion duration as a fixed hyperparameter rather than a variable inferred from semantics, inducing two systematic failure modes: duration trailing on short prompts and stage-wise semantic collapse on long, multi-stage prompts. To address this issue, we propose TPPMG (Temporal Planning-driven Progressive Motion Generation), which factorizes generation as semantics -> temporal structure -> motion realization. TPPMG comprises two components: (1) a Temporal Planner that maps text to an explicit stage-duration script via an LLM with a lightweight duration calibrator; and (2) a Progressive Diffusion Generator that bridges fixed-length diffusion training and variable-length inference through heterogeneous noise scheduling and a sliding-window mechanism, enabling budget-driven variable-length motion synthesis. To evaluate both failure modes with verifiable ground truth, we construct three benchmarks: HumanML3D-Short, HumanML3D-Concat, and BABEL-Concat, along with a comprehensive evaluation suite. Experiments show that TPPMG suppresses duration trailing (UR: 0.52 -> 0.85), substantially improves multi-stage structure recovery (PhaseRecall: 0.56 -> 0.90, BoundaryF1: 0.28 -> 0.56), and reduces boundary discontinuity at stage transitions.