RoboLab — Robotic Data Collection and Teleoperation Platform for PAL TIAGo in NVIDIA Isaac Sim (euROBIN MS3, INRIA-reviewed)

RoboLab is a Next.js + Prisma platform for collecting diverse manipulation demonstrations on the PAL TIAGo robot inside NVIDIA Isaac Sim. It supports VR teleoperation, autonomous MoveIt 2 task execution, and a sim-to-real bridge that re-uses the same intent layer on the real TIAGo.

Architecture: Web UI (Next.js 16, Prisma, SQLite) ↔ MoveIt 2 (MoveGroup, Intent Bridge with 41 high-level intents, FollowJointTrajectory Proxy) ↔ Isaac Sim (data_collector, Replicator, Articulation, multi-camera) → HDF5 export → ML training.

Features: - 41 intents covering manipulation (pick from sink/fridge, open/close fridge, bimanual pick), navigation (forward/back/rotate, nav-to-table, -fridge, -sink) and combined nav-and-pick. - Multi-camera capture (head, wrist, external) with synchronised RGB / depth / pointcloud / semantic streams. - VR teleoperation via OpenVR / SteamVR with ROS 2 Servo (Vive Pro 2 ready). - Sim-to-real bridge: same intent layer on real TIAGo. - Two procedural scenes built from YAML configs (Kitchen Fixed, Office Fixed) plus 5 task configurations (banana → sink, mug → fridge, fridge open/close, mug rearrange, office mug → desk). - 49 graspable USD assets across 8 categories (30 in-house USDA + 19 YCB benchmark objects). - Batch collection: PowerShell scripts for queued runs, HDF5 export ready for downstream policy learning. - Web UI for episode lifecycle (create, start/stop, live monitoring with SSE log streaming, sync), scene/launch-profile/object-set management. - Regression suite: 86/86 checks pass on experiments 1–3 (banana-to-sink, mug-rearrange, fridge open/close).

PAL TIAGo URDF is sourced from PAL Robotics' publicly available distribution. Object assets are 19 YCB benchmark + 30 in-house USDA; scenes are built procedurally in-house from YAML configurations.

This archive captures the v1.0.1 release, delivered for Milestone 3 of the euROBIN 3rd Open Call Technology Exchange Programme (Sub-Grant Agreement euROBIN_3OC_8). Hosting Technical Partner: INRIA (Serena Ivaldi). The MS3 deliverable was reviewed and approved by INRIA on 30 April 2026.

The accompanying SyntetiQ stack components are archived on Zenodo under DOIs 10.5281/zenodo.19818727 (DaaS portal) and 10.5281/zenodo.19818731 (IsaacSimApp). The continuously updated source is on GitHub at https://github.com/syntetiq/robolab.

Released under the Apache License 2.0.

Funded by the European Union — euROBIN, the European Robotics and AI Network (Horizon Europe Grant Agreement 101070596).