Masked Generative Motion Planning with Geometry-Guided Token Search

Lipeng Zhuang, Yingdong Ru, Shiyu Fan, Edmond S. L. Ho, Gerardo Aragon Camarasa, Paul Henderson · N/A · 2026

Summary

Abstract Summary

Generative motion planners typically use learned trajectory priors for initial generation, while leaving test-time repair to local continuous refinement. We introduce Masked Generative Motion Planning (MGMP), which extends the learned prior from efficient parallel generation to structural repair. A masked generative transformer generates discrete trajectory candidates in parallel, and Geometry-Guided Token Search (GGTS) uses scene geometry to target where to edit and which prior-supported alternatives to evaluate. This turns refinement into an efficient search over discrete motion alternatives, enabling route-level restructuring beyond local trajectory deformation. MGMP achieves 96% success on Ring Maze and 82% repair success on Controlled Route Invalidation on Kuka, exceeding the strongest external baselines by 23 and 25 percentage points, respectively. It further generalizes to unseen layouts, additional obstacles, unseen geometries, single- and dual-arm planning, and real-world Baxter tasks.

Key Points

Generative motion planners typically use learned trajectory priors for initial generation, while leaving test-time repair to local continuous refinement.
We introduce Masked Generative Motion Planning (MGMP), which extends the learned prior from efficient parallel generation to structural repair.
A masked generative transformer generates discrete trajectory candidates in parallel, and Geometry-Guided Token Search (GGTS) uses scene geometry to target where to edit and which prior-supported alternatives to evaluate.
This turns refinement into an efficient search over discrete motion alternatives, enabling route-level restructuring beyond local trajectory deformation.
MGMP achieves 96% success on Ring Maze and 82% repair success on Controlled Route Invalidation on Kuka, exceeding the strongest external baselines by 23 and 25 percentage points, respectively.

Methodology

We formulate motion planning as conditional trajectory generation. Given a start state $c_{s}\in\mathbb{R}^{d}$, a goal state $c_{g}\in\mathbb{R}^{d}$, and environment constraints $w$, the planner is required to generate feasible motion trajectories $\left\{x_{t}\right\}_{t=0}^{T-1}$.

Here $d$ is the dimensionality of the robot’s state space, $x_{t}$ is its configuration at frame $t$, and $T$ is the horizon of the full trajectory.

Figure 2: Overview of GGTS. We use geometric attribution to select edit positions, first-order scores to screen token replacements, and sequential verification to accept improving candidates, followed by optional continuous polish.

Experiments

We organize the evaluation from controlled analysis of structural repair to broader generalization and real-world execution. Ring Maze tests whether MGMP can escape local planning basins that require substantial route changes (Sec. 4.1.1 ), while Controlled Route Invalidation on Kuka isolates GGTS-based repair from initial-generation quality by initializing all methods from the same invalidated trajectory (Sec.

Sources and Demonstrations

Figures are reproduced from Zhuang et al. See the full paper for experimental details and the project page, when the paper links one, for demonstrations and videos.