[OVPHYSX] FrameTransformer sensor

[AntoineRichard]

Summary

Implement `FrameTransformer` and `FrameTransformerData` for the OVPhysX backend, satisfying the `BaseFrameTransformer` and `BaseFrameTransformerData` contracts. Computes relative transforms between a source frame and multiple target frames attached to rigid bodies.

Validation environment: `Isaac-Lift-Cube-Franka-v0` — tracks Franka end-effector pose relative to robot base.

Scope

• `frame_transformer.py` (~500 lines), `frame_transformer_data.py` (~150 lines), `kernels.py` (~100 lines)
• Complex initialization (frame resolution, index mapping, offset precomputation)
• Simple update (single transform fetch + kernel launch)

OVPhysX API Question

How to read transforms for arbitrary bodies (not just root) from ovphysx? The Articulation's `ARTICULATION_BODY_POSE` binding returns all link poses — FrameTransformer needs to index into specific links. @marcodiiga to clarify.

Dependencies

• Requires PR ovphysx backend integration #4852 merged
• Independent of other sensor specs

Spec

Full design spec: `docs/superpowers/specs/2026-04-20-ovphysx-frame-transformer-design.md`

Parent issue: #5315

---

📋 Full Design Spec (click to expand)

OVPhysX FrameTransformer Sensor — Design Spec

Issue: #5320 — [OVPHYSX] FrameTransformer sensor
Date: 2026-04-20
Status: Draft

Summary

Implement FrameTransformer and FrameTransformerData for the OVPhysX backend, satisfying the BaseFrameTransformer and BaseFrameTransformerData contracts. Computes relative transforms between a source frame and multiple target frames attached to rigid bodies. Only requires body transforms (no velocities or accelerations).

Validation environment: Isaac-Lift-Cube-Franka-v0 (manager-based) — uses FrameTransformer to track end-effector pose relative to robot base. Also validates #5322 (manipulation support).

Contract to Satisfy

BaseFrameTransformer

Inherits SensorBase. Additional abstract:

Property/Method	Type	Description
data	BaseFrameTransformerData	Transform data
num_bodies	int	Number of target frames (deprecated)
body_names	list[str]	Target frame names (deprecated)
_initialize_impl()	method	Resolve frames, build index maps
_update_buffers_impl(env_mask)	method	Fetch transforms, compute relative poses

BaseFrameTransformerData

Property	Shape	Type	Description
target_frame_names	—	list[str]	Names of target frames
target_pose_source	(N, M)	wp.transformf	Target pose relative to source
target_pos_source	(N, M)	wp.vec3f	Target position relative to source
target_quat_source	(N, M)	wp.quatf	Target orientation relative to source
target_pose_w	(N, M)	wp.transformf	Target pose in world frame (with offset)
target_pos_w	(N, M)	wp.vec3f	Target world position
target_quat_w	(N, M)	wp.quatf	Target world orientation
source_pose_w	(N,)	wp.transformf	Source pose in world frame (with offset)
source_pos_w	(N,)	wp.vec3f	Source world position
source_quat_w	(N,)	wp.quatf	Source world orientation

Where N=num_envs, M=num_target_frames.

Architecture

File Layout

source/isaaclab_ovphysx/isaaclab_ovphysx/sensors/
└── frame_transformer/
    ├── __init__.py
    ├── __init__.pyi
    ├── frame_transformer.py       (~500 lines)
    ├── frame_transformer_data.py  (~150 lines)
    └── kernels.py                 (~100 lines)

Implementation Pattern

frame_transformer.py:

1. _initialize_impl() (most complex part):

   • Resolve source prim path and verify rigid body exists
   • Resolve all target frame prim paths and verify rigid bodies
   • Compute fixed offsets for source and target frames (once)
   • Build body tracking: create a combined view of all tracked rigid bodies
   • Build per-env index arrays:
     • _source_raw_indices: maps env → position in raw transform array
     • _target_raw_indices: maps (env, frame) → position in raw transform array
   • Pre-compute offset arrays as GPU warp arrays
   • Allocate output buffers

2. _update_buffers_impl(env_mask):

   • Read all tracked body transforms from ovphysx bindings (single batched read)
   • Launch frame_transformer_update_kernel with 2D grid (num_envs, num_target_frames):
     • Fetch source transform by index, apply source offset
     • Fetch target transform by index, apply target offset
     • Compute relative: target_relative = source_inverse * target_offset
     • Store 6 outputs (source and target poses in world + relative)

OVPhysX API Requirements

API	Purpose	Notes
Body transforms	Rigid body poses for all tracked frames	Available via existing bindings

No new ovphysx APIs needed — only body transforms are required, which are already available through existing Articulation and (future) RigidObject bindings.

The key implementation question is how to create a single batched view spanning multiple distinct body paths. In PhysX, RigidBodyView accepts a glob expression. For ovphysx, the approach depends on whether TensorBindings can be queried by body index within an articulation, or whether a separate mechanism is needed.

Question for @marcodiiga: How to read transforms for arbitrary bodies (not just the root) from ovphysx? The Articulation's ARTICULATION_BODY_POSE binding returns all link poses — FrameTransformer needs to index into specific links by name.

Warp Kernels

Kernel	Purpose
frame_transformer_update_kernel	2D launch (env, frame): offset application + relative transform computation

Tests

Source: source/isaaclab_physx/test/sensors/test_frame_transformer.py
Target: source/isaaclab_ovphysx/test/sensors/test_frame_transformer.py

Validation

Isaac-Lift-Cube-Franka-v0:

• FrameTransformer tracks Franka end-effector relative to base
• Verify source/target poses match expected robot kinematics
• Verify relative transforms are consistent with forward kinematics

Dependencies

• Requires PR ovphysx backend integration #4852 merged
• No new ovphysx APIs needed (uses body transforms from Articulation bindings)
• Independent of other sensor specs — can be implemented in parallel

Estimated Scope

• frame_transformer.py: ~500 lines (initialization is complex, update is simple)
• frame_transformer_data.py: ~150 lines
• kernels.py: ~100 lines
• Tests: ~200 lines of adaptations