---
title: 'SimWeaver-Syn: Deterministic Trajectory Synthesis'
url: https://www.emergentmind.com/topics/simweaver-syn
type: topic
---

# SimWeaver-Syn: Deterministic Trajectory Synthesis

SimWeaver-Syn is a deterministic, topology-aware trajectory synthesis module within the SimWeaver pipeline, designed to generate high-quality demonstrations for deformable-object manipulation tasks entirely in simulation. It operates without the use of teleoperation, learned generative models, or post-hoc filtering, producing directly usable, closed-loop-verified demonstrations from a single seed and a canonical mesh. The module targets efficient, generalizable, and replayable demonstration generation for a range of objects, including cloths, structured garments, and bags, addressing key challenges in sim-to-real transfer for deformable manipulation [2606.15338].

## 1. Design Goals and Motivations

SimWeaver-Syn is developed with the following primary objectives:

- **Zero-teleoperation, zero-filtering**: All demonstrations are synthesized algorithmically, dispensing with human-in-the-loop teleoperation or after-the-fact discriminator-based filtering. Input is restricted to a labeled canonical mesh and a single random seed.
- **Topology awareness**: Deformable objects lack fixed frames; thus, semantic manipulation must be encoded through mesh topology. All grasps and motion planning are defined in terms of relations between canonical mesh landmarks rather than discrete keypoints.
- **Determinism**: Both grasp selection and trajectory synthesis are purely deterministic functions of the simulator state and mesh labeling. This guarantees consistent, high-throughput data generation and reliable sim-to-real policy transfer.
- **Generality**: The same synthesis protocol applies to all supported deformable assets—including rectangular cloths, garments, and plastic bags—differing only in a one-time annotation of mesh landmarks, with no per-task calibration.

## 2. Inputs and Outputs

**Inputs** required by SimWeaver-Syn:

- Canonical mesh $M$ annotated with a set of semantic landmarks $V$ (e.g., corners for cloths, sleeve tips for garments).
- Topology adjacency graph $G = (V, E)$, encoding valid bimanual grasp pairs per asset. For regular shapes, $G$ may be autogenerated; for others, labeled per asset.
- Real-time observation $\mathrm{obs}_t = \{p_v \mid v \in V\}$ from SimWeaver-Sim, where $p_v \in \mathbb{R}^3$ are 3D positions.
- Task specification (goal predicate and stage definitions per manipulation task).

**Outputs**:

- A deterministic, closed-loop-verified trajectory $\tau$, structured as a sequence of joint-space actions (e.g., $q(0 \dots T)$, gripper signals), serialized in 50-step segments at 25 Hz for use in offline policy training.

## 3. Algorithmic Pipeline

SimWeaver-Syn employs a closed-loop pipeline interleaving three deterministic components each episode:

**a) Topology-Adjacency Selection (tas):**
All semantically valid landmark pairs for bimanual grasps form a graph $G = (V, E)$. At each planning cycle, a feasibility set $F(\mathrm{obs}_t) \subseteq E$ is computed using closed-form geometric predicates, ensuring each candidate grasp pair is physically realizable. Each feasible edge is scored by a task-specific function $S(u, v; \mathrm{obs}_t)$. The selected grasp pair is:
$$(l^*, r^*) = \arg\max_{(u, v) \in F(\mathrm{obs}_t)} S(u, v; \mathrm{obs}_t)$$

**b) Topology-Aware Feasibility Predicates:**
For $(u, v) \in V \times V$, the following are enforced:

1. *Reachability*: $\min(\|p_u - b_l\|, \|p_u - b_r\|) \leq R_{ws}$
2. *Anti-cross-arm*: $\langle p_u - p_v, b_l - b_r \rangle \geq 0$
3. *Safety margin*: $\|p_u - p_v\| \geq d_{min}$
4. *Surface exposure*: For the convex hull $H_{xy}(\mathrm{obs}_t)$, $\min_{q \in H_{xy}} \|p_u^{xy} - q\| < d_{hull}$
5. *Occlusion-above*: 
   $$\mathrm{Occ}(u): \neg\, \exists v \in V \setminus G_\delta(u) \left( \|p_v^{xy} - p_u^{xy}\| < r_c \land z_v - z_u \in [\delta_z, h] \right)$$
6. *Layer separation*: $\max_{i \in \{x, y\}} \left( \max_{v \in N(u)} c_i(v) - \min_{v \in N(u)} c_i(v) \right) \leq \tau_{spread}$

Where $G_\delta(u)$ is the $\delta$-radius geodesic mesh neighborhood, $N(u)$ the local XY-neighborhood, and $c_i(v)$ the canonical coordinate.

The final feasibility set:
$$
F(\mathrm{obs}_t) = \{(u, v) \in V \times V \mid (u, v) \in E \land \bigwedge_{a \in \{u,v\}} [\mathrm{Reach}(a) \land \mathrm{Surf}(a) \land \mathrm{Occ}(a) \land \mathrm{Layer}(a)] \land \mathrm{Cross}(u, v) \land \mathrm{Safe}(u, v)\}
$$

**c) Deterministic Motion Planning and Closed-Loop Verification:**
For each attempt:
1. Observe $\mathrm{obs}_t$
2. Select optimal grasp pair $(l^*, r^*)$
3. Synthesize a motion trajectory $\tau$ to the grasp and execute post-grasp skill (e.g., fold, lift) using planners such as TOPP-RA and Pinocchio/cuRobo for inverse kinematics. The time-optimal trajectory objective:
   $$
   \min_{q(t)} \int_0^T 1\,dt \quad \text{s.t.} \quad q(0) = q_{\mathrm{start}},\; q(T) = q_{\mathrm{goal}},\; \dot{q}(t) \leq \dot{q}_{\max},\; \ddot{q}(t) \leq \ddot{q}_{\max}
   $$
4. Verify grasp and goal predicate post-execution.
5. If unsuccessful, anchor already-verified grasp and retry with updated $\mathrm{obs}_t$.

All decision points and trajectory synthesis steps are deterministic given inputs, supporting replay and robust correction of failed or ambiguous episodes.

## 4. Interface and Integration in the SimWeaver Pipeline

SimWeaver-Syn interacts as follows with the other SimWeaver modules:

| Module             | Role               | Data Exchanged                   |
|--------------------|--------------------|----------------------------------|
| SimWeaver-Asset    | Supplies asset mesh and semantic labeling | Canonical mesh $M$, landmarks $V$, adjacency $E$ |
| SimWeaver-Sim      | Simulation, observation, reset, execution | $\mathrm{obs}_t$, low-level execution, image/RGB, control |
| SimWeaver-Real     | Downstream data augmentation and learning | Demonstration trajectories $\tau$ |

Demonstrations synthesized by SimWeaver-Syn are serialized in LeRobot format, consumed by SimWeaver-Real for ISP-aware photometric augmentation, domain randomization, and VLA policy training. The workflow is strictly unidirectional: Asset $\rightarrow$ Sim $\rightarrow$ Syn $\rightarrow$ Real $\rightarrow$ Deployment, with sim-level closed-loop checks at each stage.

## 5. Empirical Performance and Ablation

**Trajectory Synthesis Quality:**

- For "T-shirt flatten" ($n = 300$): pass rate $89.7\%$, with $100/100$ replay determinism; $10.3\%$ lift failure, $0\%$ stall failure.
- *Ablations*:
  - Without closed-loop retry: pass reduces to $65.0\%$ (all additional failures are lift-failures).
  - Without topology-adjacency selection: further reduction to $56.3\%$ pass.
- Cross-method comparison ("T-shirt fold", $n = 100$): SimWeaver-Syn $97.2\%$ pass, $100/100$ replay; SIM1 pipeline $24.0\%$ pass, only $13/100$ replay, $32\%$ stall, $43\%$ lift failure.

**Real-world zero-shot deployment (five tasks, $n=23$ trials/task):**

- Mean per-task success: $91.3\%$ ($\geq 82.6\%$ per task) using $200$ demonstration trajectories/task synthesized with SimWeaver-Syn.

**Sample-efficiency and robustness (silk grasping):**

- In-distribution: Sim-trained policy achieves $100\%$ success at $200$ demonstrations (vs. $\leq 85\%$ for real-data baselines).
- Under substantial visual distribution shifts (texture, lighting, rotation): sim-to-real policy retains $100\%$ success, while real-trained baselines drop to $13\%/70\%/9\%$.

**Cost efficiency:**

- $2,824$ usable trajectories/day on $8 \times$ RTX 4090, compared to $710$ for SIM1, $104$ for real-robot.
- Unit cost: \$0.03 per trajectory (1.1\% of real-robot cost, 37\% of SIM1 cost).

## 6. Significance and Implications

SimWeaver-Syn’s deterministic, topology-aware, closed-loop trajectory generation algorithm enables scalable and efficient data synthesis for deformable manipulation tasks, sidestepping prior dependencies on teleoperation and costly post-processing. The methodology achieves fully replayable, high-yield, and cost-effective demonstration collection, supporting reliable zero-shot sim-to-real transfer without per-task calibration. Empirical results demonstrate substantial improvements in success rates, replayability, sample efficiency, and robustness to visual domain shifts over prior methods [2606.15338].

A plausible implication is that deterministic, algorithmic demonstration synthesis using explicit topology and geometric predicates—rather than imitation or generative approaches—may become a standard paradigm for sim-to-real learning pipelines in deformable object manipulation, particularly where real-world data and extensive teleoperation are impractical.

Source: https://www.emergentmind.com/topics/simweaver-syn